Investigation on principles based accounting standards

Published: October 28, 2015 Words: 1853

The 19 accounting standard declared by Financial Accounting Standards Board (FASB) in the past 30 years have affected the quality of accounting report. Two factors impel this research. First, the inflexible of General Accepted Accounting Principles (GAAP) have created accounting omission and give rise of Sarbanes-Oxley Act 2002 (SOX). FASB have granted Security and Exchange Commission (SEC) investigation on principles-based accounting standards should engage more in standard setting process.

Second, SEC had ratified registrants to file statement using International Financial Reporting Standards (IFRS) which are more principles-based and easy to apply. To understand the difference between quality and relevant attributes, this research has been conducted.

There are some indication showing that when standards contain execution guidance and exception, analyst predict failure chance will increase and insistent will diminish too. Users are able to mastery the coming standard setting trend and become profitably by understanding how the accounting standard will bring effect on the quality of prepared accounting information.

2 Motivations and Hypothesis

Previously, there was a lot of research about the improvement of accounting quality. These studies were considered as the economic consequences of personal accounting standards but all of these researches did not study on whether the accounting is actually improved.

A collective evaluation of standard setting process did not reflect on the evaluation at the individual standard level. Most of the 52 standards issued in 25 years since 1980 were industry-specific applications. The evaluation of nineteen accounting standards was based on five dimensions of the SEC study. The dimensions include whether the standard increases disclosure, contains bright lines, has an asset-liability orientation, includes alternatives or exceptions and provides implementation guidance. The research shows that there were no particular trends that were observed.

The increasing of disclosure would improve the information environment as well as the accounting quality. A bias and less representative reporting might occurred after the limitation of bright lines as they would limit the transaction structuring. The balance sheet based approach would bring positive impact on the balance-sheet oriented accounting quality metrics while negative impact to the earning and income-based metrics. The comparability was reduced as a result of subsistence of completion guidance, exceptions, and alternatives.

The way of measurement of the accounting quality would be a factor in accounting quality improvement. It was showed that accounting quality metric that were based on balance sheet would improved as a result of the improvement in the measurements of assets and liabilities and worsen if accounting quality metric was based on income statement as a result of measurement errors and possible earnings volatility. The null hypothesis set by researchers was the implementation of accounting standards and accounting quality has no relationship.

3 Research design

This research design consists of measurement of accounting quality attributes and identifying accounting standard implementation dates. Accounting quality metrics is applied as a tool of measurement which is based on the investor's valuations, analysts and also the accounting characteristics. Basically, they have two perspectives which are accounting perspective and market perspective. In accounting perspective, they apply persistence, earnings response coefficients and accrual quality measure. On the other hand, balance sheet and income statement data are supported by the explanatory power of a valuation model which is used to supply a market perspective.

Analysts' forecasts

Forecast error and forecast dispersion are the first two accounting quality metric. The forecast errors and dispersion are both positively affected by the reporting of a loss, raised earnings validity, higher leverage and changes in earning. Besides that, trend variable (TREND) and indicator variable (STANDARD) also included and its functions together with the effects from the other factors which are changing over time and may affect accounting quality and the implementation of the overall accounting standard. Analysts' forecast model:

LNERRORi,t (LNDISPi,t) = α0 + α1 LOSSi,t + α2 ANALYSTSi,t +

α3 LNASSETSi,t + α4 EPS_VOLi,t + α5 LEVERAGEi,t + α6 %ΔEARNi,t +

α7 TRENDt + α8 STANDARDt + εi,t

Persistence

Earnings persistence is estimated by regressing earnings in the current period with the earnings in the prior periods, besides that, it also captures the permanence of earnings from one period to the next. Higher earnings persistence reflects the characteristic of a higher accounting quality. Market-to-book (MB) is used to identify the relationship between high growth firm and the incremental effect on persistence. Below is the persistence model:

EARNi,t = α0 + α1 EARNi,t-1 + α2 MBi,t + α3 ( EARNi,t-1 * MBi,t ) +

α4 TRENDt + α5 STANDARDt + α6 (EARNi,t-1 * STANDARDt) + εi,t

Earnings response coefficients (ERC)

ERC is defined as the coefficient of the slope of the regression of returns (or abnormal return) on changes in earnings (or unexpected earnings). They calculate the returns as annual shareholder returns (RETURN) and earnings as the change in earning during the year (EARN). Below is ERC model:

RETURNi,t = α0 + α1 ΔEARNi,t + α2 TRENDt + α3 STANDARDt + α4 (ΔEARNi,t * STANDARDt) + εi,t

Accrual Quality

Based on Dichow and Dichev (2002) and Francis et al. (2005), they expect that the accrual quality is negatively affected with greater cash flow variability, smaller firms, reporting of loses and longer operating cycles, So, they develop an estimating equation that consists of known determinants for accrual quality. Below is accrual quality model:

SD_AQi,t = α0 + α1 LNASSETSi,t + α2 SD_CFOi,t + α3 SD_SALESi,t + α4 LNCYLCEi,t + α5 LOSSi,t + α6 TRENDt + α7 STANDARDt + εi,t

Valuation

They follow a similar approach which investigated by Collins et al. (1997) about the over-time changed in incremental explanatory power and have estimated three valuation models which are the earnings model, the book value of equity model, and also the model which incorporating both variables. Below are the three models:

PRCi,t = γ0 + γ1 BVPSi,t + εi,t (6)

PRCi,t = γ0 + γ2 EPSi,t + εi,t (7)

PRCi,t = γ0 + γ1 BVPSi,t + γ2 EPSi,t + εi,t

The explanatory powers from these three equations which have been mentioned above are needed in order to estimate and compare each of the standards after and before the implementation.

4 Samples and data

The samples which were used in this research were within 29 years span dated from 1976 till 2005. Observation which was non-December year-ends and data which were lack lagged were removed. Mean and median statistics were done and difference between the samples is recorded. Smaller and less firms were included when the sample sizes increases. The forecast sample shows that it is occupied more by larger firms. All samples show positive earnings from the median data. In each sample, 50% of the observations are reported as positive earnings four of the five samples show the negative earnings.

5 Results

There was more than one accounting standard that was implemented each year. The total of 11 standard setting events was compiled in order to investigate it. In accounting standard variables, the estimated coefficient was the changes of the accounting quality metric after the implementation of the accounting standard with controlled variables. The mean of the coefficient on the accounting standard event is tested to determine its significance towards 0. Limitation of years done is used to reduce the sampling bias and strengthen the generalizability of the results. The sample period is the 4 year period that ends 2 years before and 4 year period that begins 2 years after the implementation of the standard. To avoid adoption year effects, exclusion of the implementation year, former year and subsequent year was done. The implementation periods allow the effects of the non-accounting effects towards the tests as each year data are variously relevant to various standards.

In analyst forecast-based, the forecast errors decreases as new standards are implemented along with new information on the environment in the forecast earnings analysis. For forecast dispersion, there were no differences before or after the implementation of the new standards. In accounting-based, it measures persistency, ERCs and accrual quality. For the persistency, there were no differences before or after the implementation of the new standards. For ERC and accrual quality, both decreases which mean that the accounting qualities had been lowered. However, the relevancy towards the balance sheet increases with the consistency trend of the asset and liability that focus on accounting standards.

In valuation measurements, consideration is made towards the explanatory power of book value per share (BVPS), earning per share (EPS) and combination valuation models. There was two analysis method was conducted. Firstly, comparison between the explanatory power before and after to each new standards is performed using all three models. For BVPS and combined models, there were no differences before or after the implementation of the new standards. For EPS model, significant decrease is discovered in the explanatory power after the implementation with the consistency trend of the asset and liability that focus on accounting standards. Secondly, using the Collins approach, there are three sources of the combined adjusted R² that is the factors common to both BVPS and EPS models; unique to BVPS model; and unique to EPS model to find out the differences before and after the new standards are implement. No differences occurred in explanatory power for common source, EPS source or combined model. However, there is a significant increase in explanatory power towards the BVPS model, consistently with the focuses on the balance sheet on accounting standards.

To determine the firmness of the results, sensitivity tests are used. Generally, all the results and inferences are firm except the marginal accrual quality that is not significant in the following case. Consideration is made during the process of taking the samples and measuring the variables. Assumption is made that the dispersion is equal to zero. Only firms that are mostly impacted by the standard are taken into account. Sample is expended for each standard. Potential outliers of the independent variables are winsorized and re-estimated. ERC model is taken based on the level of earnings instead of change in earnings. Prior research suggested that additional ERC variable should be included as the main effects and loss firm to be evaluated differently than profit firms by estimating ERC and valuation model separately for both firms; and include the firm's book values and EPS interactions and also the firm's indicator in valuation models. ERC is weakened but still consistent with those reported. Differentiation is made on the attributes that has good or bad influence on the accounting quality. So, estimated coefficient is based on the classification of the accounting standard on the subject of increased disclosure, income statement, bright lines and implemented guidance, exception or alternatives. Evidence gained is qualified as the classifications are necessarily critical and limited due to comparison of only 11 coefficients.

6 Conclusion

This research had examined 11 general accounting standard setting events in 19 accounting standard regarding accounting quality from 1976 to 2005. It has provided verifiable evidence in accounting standard setting process and it was generally comply with FASB's asset-liability approach. The result of this study are generally granted FASB's new standard, which concentrate on balance sheet (asset-liability) and expand disclosures. The explanations and present of the results is subject to accounting quality standard and shall be substitute in particular companies.