International Journal of Academic Research in Accounting, Finance and Management Sciences

Vol. 3, No.3, July 2013, pp. 234–241

ISSN: 2225-8329
© 2013 HRMARS
www.hrmars.com

Income Smoothing and the Cost of Debt and Credit Ratings

Abdolkarim MOGHADAM1
Mehdi BAHARMOGHADAM2
Mojtaba MOHAMMADZADEH3
1
Payame Noor University, Tehran, Iran
2
Shahid Bahonar University, Kerman, Iran
3
Kerman Science and Research Branch, Islamic Azad University, Kerman, Iran

Abstract The goal of this research is to examine the effect of income smoothing on the cost of debt and the
credit rating. The statistic community is the accepted firms in Tehran’s stock exchange during 1385-
1389, that the statistic sample has chosen from them. Also in this research we used Jones’ modified
model for measuring the discretionary accruals (measure of income smoothing). And we used
regression analysis for testing the research hypothesis. The results of this research show that there is
significant and negative relation between income smoothing and cost of debt. And also there is
significant and direct relation between income smoothing and firm’s credit rating.
Key words Income smoothing, cost of debt, credit rating, discretionary accruals

DOI: 10.6007/IJARAFMS/v3-i3/133 URL: http://dx.doi.org/10.6007/IJARAFMS/v3-i3/133

1. Introduction
In recent years, earning management has increased many serious questions for market financial
regulators, investors and academic researches of many advanced countries and also the prevalence of earning
management between companies is momentous for legislators and professional persons. Investors, creditor
and financial analysts are interested in having more information concerning income smoothing in invested
firms especially if this action is effective on risk and return.
Always earning manipulation doesn’t show more than real profit figure ,but sometimes managers
prefer to report profit figure less than real for reduction of accountability against some body that want a
response. And by this way don’t allow sending out the addition of cash because of tax or dividend. Some
believe that there is no difference between earning management and dishonesty but the truth is that earning
management is done in framework of accepted accounting principles. And managers in addition to
compliance of accounting standards manage the earning (Scott, 2003). In this research the effect of income
smoothing has examined on cost of debt and credit rating by Tucker and Zarowin’s model and also the
research questions are:
1. What is the effect of income smoothing on cost of debt?
2. What is the effect of income smoothing on credit rating?
Although there are many researches about income smoothing but there is no research about the effect
of income smoothing on cost of debt and credit rating. So this research is new.

1.1. Definition of income smoothing

There are various definitions of income smoothing and every researcher based on his/her method that
used in income smoothing has offered one definition. Barna and his colleagues know income smoothing as
intentional reduction of fluctuation within accounting principles to the extent it seems normal to companies.
Byldman know income smoothing as efforts made by management to reduce abnormal changes in income
smoothing and within accounting principles. Both definitions are emphasizing on earning manipulation based
International Journal of Academic Research in Accounting, Finance and Management Sciences
Vol. 3 (3), pp. 234–241, © 2013 HRMARS

on profit of past years. Such that the abnormal return is reduced and reported earning be the same with
expected earning.

1.2. Credit rating

Credit rating is increased with both positive firm’s image in society and indicator‘s improvement and
firm’s financial ratios. So numerator of brand value equation and at last firm’s brand value is increased. High
firm’s market share means more awareness and loyalty of customers. By increasing market share, firm’s
financial capability in providing advertising budget be increased. And also financial ratios are consolidated.
These two processes make attractiveness of firm’s shares in market be increased and also increase value of
company’s shares in stock exchange. Also this process makes credit rating be consolidated and facilitate
financial providing by short and long term debt with lower rate (Azizi et. al, 1390).

2. Literature review
Pourheydari and Aflatouni (1385) have examined the motivations of income smoothing in accepted
firms in Tehran’s stock exchange with using of discretionary accruals. The results of this research show that
income smoothing is done with discretionary accruals by managers of Iranian firms. And income tax and
deviation in operating activities with discretionary accruals are principal stimulus for income smooth and
against western researcher’s results, the firm’s size, the ratio of debt to total assets (debt deals) and earning
fluctuating have little importance.
Molanazari and Karimi (1386) are examined the relation of income smoothing with firm’s size and the
kind of industry in accepted firms in Tehran’s stock exchange. The results of this research show that there is
strong correlation between firm’s size (sale) and income smoothing. And this correlation is reverse
correlation. Also there is no significant difference between earning smoother firms from an industrial point
of view (axial or circumferential), but there is just a weak relation between industry and income smoothing in
gross profit level.
Hashemi and Samadi (1388) have examined the effect of income smoothing on information content in
accepted firms in Tehran’s stock exchange the results of examining hypothesis show that income smoothing
increase profit ability in predicting it and future operating cash flows while profit ability isn’t increased in
predicting accruals by income smoothing.
Shorozi and Pahlavan (1389) have examined the effect of firm’s size on income smoothing. In this
research the effect of firm’s size on income smoothing is examined in accepted firms in Tehran’s stock
exchange and also 352 accepted firms in Tehran’s stock exchange are examined during 1381-1385. The results
of this research show that there is positive relation between firm’s size and income smoothing.
Demory et. al. (1390) have examined the relation between income smoothing ,quality of earnings and
value of firm in accepted firms in Tehran’s stock exchange the results of multiple regression show that
investors are price most value for qualified earnings smoother firms and least value for no qualified and no
earnings smoother firms.
Moses (1987) and Ndubize, G.A. and Tsetsekos (1991) argue that firms begin to smooth for reduction of
risk. In other words the lack of considerable fluctuation in earnings make sure the creditors that business unit
can pay their demands in future.
Tucker and Zarowin(2006) have examined income smoothing on its information content . Based on the
results of their research, income smoothing increase income’s information content and smoothed earnings
present information about earnings, cash flow and future accruals.
Tseng,L,J. and Chien, W.L. (2007) resulted that there is strong negative relation between profitability
and income smoothing. Also in their research four factors such as profitability, debt quantity, quantity of paid
earnings and firm’s size are introduced as motivations for income smoothing.
Huang et al. (2009) have examined the effect of artificial smoothing and real smoothing on value of
firm. The results of their research show that the value of firm is decreased because of using unusual deferred.
And as a result using of real smoothing becomes increase.

235
 International Journal of Academic Research in Accounting, Finance and Management Sciences
Vol. 3 (3), pp. 234–241, © 2013 HRMARS

3. Research conceptual model

The research conceptual model is Tucker and Zarowin’s model (2006).
Income smoothing is estimated based on negative correlation between changes in firm’s discretionary
accruals (ΔDAP) and changes in optional benefit (ΔPDI). The Jon’s model (1991) is used for estimating the
discretionary accruals. That in it:

= (1/ )+ + + + (1)

ASSETS is total assets

ACCRUALS is total accruals
ΔSALES is changes in sales

PPE is property, plant and equipment and ROA is return on assets using net income over lagged total
assets. ACCRUALS, ∆SALES and ∆PPE are each deflated by the beginning-of-year total assets (ASSETS). Non-
discretionary accruals (NDAP) are represented by the fitted values of regression 1:

= (1/ )+ + + (2)

And discretionary accruals (DAP) are represented by the deviations of actual accruals from NDAP:

DAP = ACCRUALS- NDAP )3(

The managed income series (PDI) is calculated as net income minus discretionary accruals, or

PDI = NI – DAP (4)

The TZ statistic is the correlation between the change in discretionary accruals and the change in un-
managed income, i.e., Corr (∆DAP,∆PDI), using the current year’s and past four years’ observations. Firms with
more negative correlations are higher smoothing firms.

3.1. Research hypothesis

This research is based on following hypothesis that after doing research and analyzing its results we
experiment them and know their truth and untruth with statistics testing in significant level.
1. Income smoothing decreases cost of debt.
2 .Income smoothing increases credit rating.

3.2. Community and statistical sample

The goal of this research is examine the relation of income smoothing with cost of debt and credit
rating. Based on necessary condition, I proceed to review and extract data from financial statements of
sample firms. The restricted area of this research is accepted firms in Tehran’s stock exchange. The statistical
sample has chosen based on the following condition that with systematic elimination, the number of samples
for testing of first stage includes 399 firms and for testing of second stage includes 314 firms.
1. Between years 1385-1389 have participated in exchange.
2. Their data bank be complete from 1385 -1389 and don’t have data limitation.
3. Their financial year terminate to March.

3.3. Data and descriptive statistics of research:

A. Testing of first stage
The first equation descriptive statistics of estimating discretionary accruals are in table 1. That includes
number of observation, mean, median, standard deviation, minimum and maximum. Mentioned quantities
show research data.
236
International Journal of Academic Research in Accounting, Finance and Management Sciences
Vol. 3 (3), pp. 234–241, © 2013 HRMARS

Table 1. Research variables descriptive statistics is related to first estimating of discretionary accruals

ROA PPE SALES ASSET ACCRUAL

6/393189 5/420920 89628/32 0.00000588 -37221/55 mean
7/516577 50/60173 161779/9 0.00000273 5/6325- Median
6115/103 25165557 97/11338 000129/0 9457277 Maximum
207/138- 000000./0 0504/390- 0.0000000052 -23535289 Minimum
06695/20 1705797 4948/336 0.0000105 5/877151 Standard deviation
1938 1938 1938 1938 1938 Number of observation

Above information has extract from research sample that include 399 firms. Number of observation
consists of five years existing data of sample firms. Manner of choosing sample is explained in previous
chapter.

B. Testing of second stage

Table 2 and 3 show summary descriptive statistics of second stage‘s test i.e. examining cost of debt that
include 314 firms. Because of information’s defect, the number of observation is 310 companies.

Table 2. Descriptive statistic

Standard deviation mean Maximum Minimum Number of observation

556000 -35781/515 2969117 -6272735 399 ACCRUAL
469000 5837/4889 51/5831363 -1305385/2 335 DAP
1642867/0 -0/9155193 09719/0- -1 335 TZ
335 Valid N (listwise)

Table 3.

TANGIB ROA COVER MKBK DEBT SIZE IS AVEYIELD

7/432967 052288/3 058065/0 602012/1 849522/0 48461/12 522079/0 3/656295 mean
76089 632043/5 0 287334/1 688426/0 51436/12 531746/0 648694/3 Median
25165557 76097/51 1 499477/8 485414/6 24168/18 1 9/733159 Maximum
0 0611/112- 0 62975/0 085628/0 652071/7 0 29046/4- Minimum
2096137 98869/19 234244/0 997731/0 743152/0 509456/1 297154/0 742265/1 Standard
deviation
890988/9 987602/1- 7794/3 507474/3 520246/4 230558/0 08296/0- 42185/0- skewness
1312/107 64546/10 28387/15 52939/19 53549/27 021007/4 782406/1 270949/5 elongation
310 310 310 310 310 310 310 310 Number of
observation

3.4. Relative statistics

A. Estimate of discretionary accruals (first stage of testing)
The regression results of estimating discretionary accruals are in following table. Firstly I summarize and
classify the raw data of database’s file then I do necessary calculations in Excel then I enter screened data as
data panel in Eviews, data processing software, and the following regression output has obtained with
399firms and 1938 observations as table 4.

= (1/ )+ + + + (5)

237
 International Journal of Academic Research in Accounting, Finance and Management Sciences
Vol. 3 (3), pp. 234–241, © 2013 HRMARS

Table 4.

DEPENDENT VARIABLE ACCRUAL

Coefficient Std. Error t-Statistic Prob. Variable

0/143219- 09/65699 -2.179924 0.0294 C

1120000000- 2300000000 -0.484567 0.6281 ASSET
321/8866 65/28753 1.923842 0.0546 SALES
0/323828 150295/0 0.942332 0.3462 PPE
8822/656 1567/257 4.877859 0.0000 ROA
R-squared 0.427610 Mean dependent var -37221.44
Adjusted R-squared 275820/0 S.D. dependent var 877151.5
S.E. of regression 746445.2 Akaike info criterion 30.06832
Sum squared resid 853000000000000 Schwarz criterion 31.23795
Log likelihood -28729.20 Hannan-Quinn criter. 30.49847
F-statistic 2.817115 Durbin-Watson stat 2.412730
Prob (F-statistic) 0.000000

According to above table the considered regression is significant at 5% level, completely .because
Prob’s statistics (F – statistic) is under 5%.also Durbin-Watson stat is near 2 and this show that regression
variables don’t have internal correlation and one of classical regression assumptions has regarded.
By comparing the mean of obtained coefficients with Tucker and Zarowin (2006) and Sabramanim
(1996) results we understand that the mean of coefficient is 123% and have positive significant
relation with accruals and it is agreeable with TK results. Like TK study we found positive significant relation
with accruals for ROA variable that had 76.44 coefficients, with this difference that obtained coefficient by TK
is 4.16 for ROA, We can know the reason of this difference because of difference in elected sample and time
and place interval of mentioned subject.
Obtained data from regression (1) consist of estimated accruals and wastes. Estimated accruals are the
same with non-discretionary accruals (NDAP) that their equation is in the following.

= (1/ )+ + + (6)

And discretionary accruals (DAP) obtain from difference of first and second equation i.e. difference of
real and non-discretionary accruals that are wastes in regression 1’s results.

DAP = Accrualst – NDAP (7)

Unmanaged income (PDI) is net income after reduction of discretionary accruals.

PDI = NI - DAP (8)

After calculation of DAP and PDI based on five years firms, The correlation changes of these two
variables for each firm calculated in 1385-1389. Negative correlation (TZ statistic) Show income smoothing, so
whatever the negative correlation be more, the income smoothing be more too. So we choose firms with
negative correlation and rank them based on smoothing. Then we allocate number 1 to firms with maximum
of income smoothing(the most negative correlation) and number 0 to firms with minimum of income
smoothing(the least negative correlation) and we called it as IS.

B. The costs of debt (second stage of testing)

In foreign countries there are validation institutions like Andpurz standard, Fich, Modiz, etc. these
institutions rank firms based on credit and firms with better credit can take a debt with lower interest rate so
their cost of debt is lesser. And vice versa firms with lower rate can’t fund by debt Because of greater credit
risk, comfortably. So they have to pay more interest to encourage creditors to debt them. But in Iran there is
238
International Journal of Academic Research in Accounting, Finance and Management Sciences
Vol. 3 (3), pp. 234–241, © 2013 HRMARS

no such this institution. So we use from another measurement for cost of debt and credit rating. Thus we use
annual mean’s natural logarithm of cost of financial as indicator of firm’s cost of debt and credit rating.
For estimating the relation of firm’s cost of debt and credit rating with income smoothing, we use from
a sectional regression that its independent variable is annual mean’s natural logarithm of cost of financial
(AVEYIELD), this research key independent and considered variable is income smoothing rank (IS) and control
variables consist of: the variables of firm’s size(SIZE), debt ratio (DEBT), the value of stock market plus book
value of the debt to book value of the assets (MKBK), dummy variable (COVER) (if operating cash flows are
greater than current liabilities takes a value of 1, otherwise it takes a value of 0), output of assets (ROA) and
the variable of property plant and equipment (TANGIB).

AVEYIELDJ = (9)

Hypothesis confirmation
First hypothesis: income smoothing reduces cost of debt.
Second hypothesis: income smoothing Increase firm’s credit rating.
According to table5, there is significant negative relation between income smoothing (IS) and cost of
debt(Prob<.05) and the coefficient of income smoothing is -1.48. so we conclude that whatever the income
smoothing be greater ,the firm’s cost of debt be lesser because these two variables have reverse(negative)
relation. Thus the first hypothesis will confirm.

Table 5.

AVEYIELD Dependent variable

Coefficient Std. Error t-Statistic Prob. Variable
-1.487870 0.242004 -6.148119 0.000 IS
0.397418 0.017228 23.06790 0.000 SIZE
0.272171 0.215326 1.263996 0.2072 DEBT
-0.404237 0.098689 -4.096080 0.0001 MKBK
-0.958183 0.335424 -2.856628 0.0046 COVER
-0.028213 0.007297 -3.866230 0.0001 ROA
0/000000117 0/0000000363 3/235655 0/0015 TANGIB
R-squared 0.463536 Mean dependent var 3. 646294
Adjusted R-squared 0.452913 S.D. dependent var 1. 742265
S.E. of regression 1.288672 Akaike info criterion 3.367423
Sum squared resid 503.1843 Schwarz criterion 3.451797
Log likelihood -514.9505 Hannan-Quinn criter. 3.401152
Durbin-Watson stat 1. 8147795

Whatever the firm’s credit rating be greater, firm can fund by debt and with lower cost of interest easily
because of lower credit risk. Thus there is reverse relation between credit rating and firm’s cost of debt. Since
we confirm income smoothing and firm’s cost of debt have reverse relation before, we conclude that income
smoothing and firm’s credit rating have direct relation. So the second hypothesis will be confirmed too.

Control variables
Coefficient of firm’s size is 39 and coefficient of property, plant and equipment is 1.17. They have
significant and positive relation with firm’s cost of debt. This shows that if the company be greater or have
more property, plant and equipment and its cost of debt is greater too.
Coefficient of MKBK, COVER and ROA is -.41,-.96 and -.28. They have significant and negative relation
with firm’s cost of debt. This shows that firms with more MKBK, COVER and ROA have lower cost of debt and
vice versa.
But these results may not be true because of correlation between variables. So, for reduction of
correlation between variables, estimate the model as following too.

239
 International Journal of Academic Research in Accounting, Finance and Management Sciences
Vol. 3 (3), pp. 234–241, © 2013 HRMARS

AVEYIELD= (10)

The achieved results of above model presented at table 6.

Generalities
By adding variables like IS*SIZE and IS*DEBT to above model, the explanatory power of independent
variable increase from .46 to .54. And all of independent variables have significant relation with dependent
(cost of debt) variable (prob<.05). Also Durbin-Watson stat is 1.97. Compared to previous mode, it is near to 2
and this shows the lack of internal correlation between independent variables.

Table 6.

AVEYIELD Dependent Variable Variable

Coefficient Std. Error t-Statistic Prob.
-8.382510 1.148767 -7.296965 0.0000 IS
0.665194 0.100030 6.649960 0.0000 IS*SIZE
-0.715702 0.352520 -2.030246 0.0432 IS*DEBT
0.319107 0.024559 12.99325 0.0000 SIZE
0.839052 0.353282 2.375021 0.0182 DEBT
-0.234054 0.095333 -2.455118 0.0146 MKBK
-0.828776 0.315442 -2.627349 0.0090 COVER
-0.031832 0.006924 -4.597182 0.0000 ROA
0000000984/0 0000000344/0 862539/2 0045/0 TANGIB
R-squared 0.536275 Mean dependent var 3.646294
Adjusted R-squared .523950 S.D. dependent var 1.742265
S.E. of regression 1.202100 Akaike info criterion 3.234619
Sum squared resid 434.9583 Schwarz criterion 3.343101
Log likelihood -492.3660 Hannan-Quinn criter. 3.277986
Durbin-Watson stat 1.973266

According to output 6 , we understand that IS*SIZE ‘s coefficient is .67. and IS*SIZE is also positive
and significant, indicating that large firms experience more costs of debt, that this is less likely to be the case
for higher smoothing firms. Also we see in table 6 that IS*DEBT’s coefficient is -72. And IS*DEBT have negative
and significant relation with cost of debt and we can conclude that higher debt ratio and lower credit rating
have more relation with cost of debt. So the effect of income smoothing is more than debt’s ratio effect .also
we understand that debt’s ratio DEBT that its coefficient is .84 have positive and meaningful relation with
firm’s cost of debt. It means that firms with higher debt’s ratio That have higher credit risk should pay more
interest for absorption of financial resources by debt that this make to increase the costs of debt.
In addition achieved results of renewed regression (table 6) are more similar to previous results (table
5) that higher income smoothing has significant relation with lower cost of debt.

4. Conclusion
The results of this research show that there is negative relation between income smoothing and firm’s
costs of debt, so if income smoothing be high, cost of financial fund will reduce by debt. In other words
creditors at the time of granting financial facilities to firms consider its financial stability to be sure of basis
receipt and their claims’ interest. If firms have lower financial stability, they can fund just with paying more
interest by debt. Thus firm’s managers use income smoothing as one of financial stability ways to reduce its
cost of debt by increasing financial stability.
Since one of the firm’s important financial sources is financing by debt and financial costs are the major
parts of firm’s costs, so firms want to reduce their financial costs and increase their performance by condition
improvement and attract firm’s condition with the help of income smoothing. Also financial markets have
positive reaction by shares price increasing of such firms.

240
International Journal of Academic Research in Accounting, Finance and Management Sciences
Vol. 3 (3), pp. 234–241, © 2013 HRMARS

Also the results of this research show that income smoothing make to increase firm’s credit rating.
Because income smoothing reduces the firm’s credit risk by increasing the forecasts accuracy and creditors
want to grant borrowing with lower cost.
In summary research hypothesis and their results are as follow:

Results Research Hypothesis

Confirmation Income smoothing decreases cost of debt
Confirmation Income smoothing increases credit rating

References
1. Pour heydari, O and Aflatouni, A. (2006). Study of income smoothing‘s motive in accepted firms in
Tehran stock exchange. Accounting and auditing studying. Summer; 55-70: (44)13
2. Demori, D. Aref Manesh, Z. Abbasi Mosslo, Kh. (2011). Study of relation between income smoothing,
quality of income and firm’s value in accepted firms in Tehran stock exchange. Financial accounting studying.
Spring; 7 (1) 3: 39-54.
3. Shourozi, M. Pahlavan, R. (2010). The effect of company size on income smoothing. Management
studying. Winter; (87) 21: 69-80.
4. Mollanazari, M. Karimi Zand, S (2007). Studying the relation between income smoothing with firm
size and kind of industry in accepted firms in Tehran stock exchange. Accounting and auditing studying.
Spring; (47) 14: 83-100.
5. Hashemi, A. Samadi, V. (2009). The effect of income smoothing on its information content in
accepted firms in Tehran stock exchange. Accounting investigations. Spring (1) 1: 148-167.
6. Emhoff E.A. (1977). “Income Smoothing – A case for Doubt”, Accounting Journal Spring 1977.
7. Eckel, N. (1981). “Income Smoothing Hypothesis Revisited”, Abacus, June, pp. 28-40.
8. Moses, D. (1987). Income Smoothing and Incentives: Empirical Tests Using Accounting Changes. The
Accounting Review, (April), pp. 358-377.
9. Ndubize, G. A. and Tsetsekos, G. P. (1991). Financial Currency Smoothing and Managerial Incentives:
An Empirical Test. The International Journal of Accounting, No.27, pp. 27-37.
10. Pinghsun Huang, Yan Zhang, Donald R. Deis, Jacquelyn S. Mof, (2009). Does artificial income
smoothing and real income smoothing contribute to firm value equivalently? Journal of banking & Finance,
33, pp. 224-233.
11. Richhie, N. (2009). “Income smoothing and the Cost of Debt”, Wilfrid Laurer University.
12. Tseng, L.J. and Chien, W.L. (2007). The Relationship between Income Smoothing and Company
Profitability: An Empirical Study. International Journal of Management, Vol. 24, No. 4, pp. 727-733.
13. Tucker, J., and P. Zarowin (2006). Does income smoothing improve earnings informativeness.
Accounting Review 81 (1): 251-270.

241