Aljalahma, Fatema
(2022)
An investigation of accrual-based earnings management in the UK.
PhD thesis, University of Nottingham.
Abstract
This thesis aims to investigate the reliability of using accrual-based measures as a proxy for earning management in the UK. The earnings management literature has undergone immense development over the last decades, initially having focused on the establishment of abnormal accruals as a proxy for earnings management, thereafter more critical papers questioned the reliability of the previously developed measures resulting in modifications of pre-existing measures.
This thesis builds on the critique that earnings management is not as prevalent as commonly suggested and that there might be concerns about how earnings management is measured. Therefore, the main research question asks - how reliable are accrual-based measures in assessing earnings management in the UK? The following objectives have been set to answer this question: the first objective is to examine the association of abnormal accruals generated from different measures in the literature, and whether they differ from the actual accruals of a firm. The second objective is to examine the persistence of abnormal accruals generated from different measures, if earnings management are a one-time manipulation they should not be persistent. The third objective is to examine the power of accrual-based measures. The fourth objective is to examine the dependability of accrual-based measures in capturing earnings management in different scenarios.
In order to meet these objectives this thesis empirically tests the correlation, power and persistence of 37 accrual-based measures of earnings management identified in the literature. In addition, it replicates five published UK papers that use abnormal accruals as proxies for earnings management, altering the estimation of their earnings management proxies to abnormal accruals generated from the 37 different measures, as well as using placebo tests such as actual accruals, lead and lagged abnormal accruals.
The main findings of the thesis shows that abnormal accruals, which are used to proxy earnings management, generated from the 37 different measures are highly correlated to each other and to the actual accruals of the firm. These measures therefore have low power and could suffer from problems with regards to correlated omitted variables. The results of persistence show that abnormal accruals do not appear to reverse over time, which is a key assumption in earnings management. In terms of the replications, this research finds other explanations for the results which could have influenced the relationship with no direct attribution to earnings management. In other scenarios, results were not found using any of the 37 accrual-based measures of earnings management, or using the actual accruals of the firm, indicating that their relationship to accruals are not robust.
This thesis contributes to the earnings management literature in theory and practice. In particular it provides evidence that the accrual-based measures of earnings management literature are not as insightful as projected by researchers. Therefore, regardless of researchers’ use of measures in estimating abnormal accruals, whether previously established, newly developed or actual accruals, the results do not differ. This may imply that the new measures still suffer from problems in correlated omitted variables, thus do not capture earnings management. Future research should contemplate the use of specific accruals such as allowance for bad debt, deferred revenue or tax expense instead of aggregated accruals.
| Item Type: |
Thesis (University of Nottingham only)
(PhD)
|
| Supervisors: |
Chen, Jing
Stark, Andrew |
| Keywords: |
Accrual basis accounting; Earnings management |
| Subjects: |
H Social sciences > HG Finance |
| Faculties/Schools: |
UK Campuses > Faculty of Social Sciences, Law and Education > Nottingham University Business School |
| Item ID: |
69120 |
| Depositing User: |
Aljalahma, Fatema
|
| Date Deposited: |
27 Jul 2022 04:40 |
| Last Modified: |
27 Jul 2022 04:40 |
| URI: |
https://eprints.nottingham.ac.uk/id/eprint/69120 |
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
