Academy of Accounting and Financial Studies Journal (Print ISSN: 1096-3685; Online ISSN: 1528-2635)

Research Article: 2021 Vol: 25 Issue: 6

Existence of Dividend Smoothing After the Financial Crisis and the Sensitivity of Firm Value to Dividend Smoothing: Evidence from Korea

KyungJae Rhee , Geumgang University

Citation Information: Rhee, K.J. (2021). Existence of dividend smoothing after the financial crisis and the sensitivity of firm value to dividend smoothing: evidence from korea. Academy of Accounting and Financial Studies Journal, 25(6), 1-8.


This study examines whether there exist changes in dividend smoothing in Korea around the 2008 financial crisis, and analyzes the effect of dividend smoothing on firm valuation pre- and post-financial crisis. The results are as follows: First, the degree of dividend smoothing after the 2008 financial crisis has decreased by 20.5%. The market has positively reacted to smoothed dividend announcements, but the post-crisis market response diminished compared to the pre-crisis level. Second, the market shows a positive and increased (decreased) reaction for more (less) smoothing dividend announcements compared to the pre-crisis level. The results indicate that the market prefers greater dividend smoothing and that the effect of dividend signaling is reduced but still exists in Korea. Finally, a negative and significant coefficient sign of SOA for more smoothing dividend announcers indicates that dividend smoothing behavior plays an important role in the firm valuation for more smoothing dividends.


Divi dend Smoothing, Abnormal Returns, Firm Valuations, S peed Of Adjustment, Financial Crisis


Since Lintner’s (1956) work, financial economists have explored the properties of stable dividends (Alli et al., 1993; Brav et al., 2005; Fama & Babiak, 1968) arising from signaling models (Bhattacharya, 1979; Ross, 1977). Dividend signaling theory, in which firms convey private information, has developed to explain excess abnormal returns following dividend change announcements (Aharony & Swary, 1980; Bernhardt et al., 2005; Pettit, 1972). Arguments have followed over which factors may affect dividend smoothing from a perspective of information asymmetry or agency theory (DeAngelo & DeAngelo, 2007; DeMarzo & Sannikov, 2016; Guttman et al., 2010; Lambrecht & Myers, 2012; Leary & Michaely, 2011; Mahmudi & Pavlin, 2013). The empirical evidence is still inconclusive about whether dividend announcements affect firm valuation and whether management uses dividends as a signaling device.

Paying dividends is closely related to a firm’s financial status: a high level of dividends requires considerable cash. If a high level of dividends is maintained, it may have a negative impact on the liquidity of the firm because it requires a considerable dividend amount regardless of the corporate performance. On the other hand, a firm with a smooth dividend policy tends to pay stable dividends regularly. The financial crisis of 2008 halted rollover by financial institutions in developed countries, increasing volatility (Calvo Mendoza, 2000) in the financial market in Korea (Kim, 2012; Stiglitz, 2010) 2010). This financial instability has increased the importance of securing cash and affected dividend stability (Rhee & Park, 2018). TThe financial crisis led market participants to question whether the information is reliable and whether dividend announcers are financially stable. These doubts are likely to result in diminished market response to dividend announcements. Thus the reactions of the stock market are expected to be less sensitive than would have been the case in the pre financial crisis period.

This study, using cash dividend data from Korean firms, analyzes the changes of dividend smoothing behavior in Korea in the pre and post financial crisis era, examines the market reaction to dividend announcements, and analyze s the relation between dividend smoothing and firm value around the financial crisis . This article attempts to improve the literature on tendency of dividend smoothing and firm values to dividend smoothing around the financial crisis The empirical results show that the speed of adjustment (SOA) of Korean firms increased by 20. 5 % compared to the pre crisis level This indicates that the financial crisis caused Korean firms to forego smoothed dividends . T he post crisis market responses to dividend announcements are still positive but show decreased reaction compare d to the pre crisis response s , indicating the decreasing role of smoothing dividends as a signaling device in the post crisis period . A negative and significant coefficient of the SOA for firms that decreas e SOA suggests that dividend smoothing affects the value of firms with more smooth dividend

Hypothesis andMethodologies

The 2008 financial crisis reduced liquidity supply and increased the volatility of the financial market and business risk; accordingly, firms’ cash holdings have emerged as an important issue in the post-crisis era. In such an economic situation, stable and regular dividends can be a burden to management, but in the market, they can serve as a signal for the company’s financial health. In addition, Lintner (1956) finds that US firms pay stable dividends, which are received positively in the market. Fama Babiak (1968) suppor t Lintner s (1956) argument by showing a correlation between dividend and profit. On the other hand, Brav et al. (2005) point out that in the case of insufficient corporate profit, additional external capital costs may arise if dividend reduction is not achieved thro ugh dividend smoothing. The dividend smoothing policy, which is managed from a long term perspective, can be subordinated to the crisis situation. Thus, the following hypotheses are proposed:

H1-1 Market response to dividend announcements will be lower than before the financial crisis.

H1-2 The market will show increasing reaction to more smooth dividends, but decreasing reaction to less smooth dividends.

The dividend smoothing practice is based on the concepts of asymmetric information (DeMarzo & Sannikov, 2016; Lambrecht & Myers, 2012; Leary & Michaely, 2011) and the signaling aspect (Aharony & Swary, 1980; Bernhardt et al., 2005; Pettit, 1972). The advantage of dividend smoothing is that it does not convey unnecessary signals to the market by avoiding sudden dividend changes due to temporary earnings shocks. The market value of information about making dividends more smooth will be of better quality than less smooth dividends, and moreover it will be enough to reduce information asymmetry between management and outside investors. Thus, the second hypothesis is proposed proposed:

H2 Firm value is positively related with more dividend smoothing.

In order to measure the market response to the dividend announcement of the dividend smoothing firm before and after the financial crisis, the market adjusted model (MktAdj), the mean adjusted model (Mean Adj ), and the market model (MktM) are used. The three models are used to analyze market responses by supplementing the shortcomings of each model. For each model, cumulative abnormal returns (CARs) are measured from day 1 to day +1, with the dividend announcement date as the event date. The CARs a re measured using the market adjusted model of Equation (1). R i,t and R m,t are the returns of the firm the market returns.


Equation (2) measures the CARs using the mean adjusted model. Equation means the average return of firm i . In this study, the average return (30 days from 2 days to 31 days) before the event date is used.


Equation (3) measures the CARs using a market model. The market model is measured by regression analysis of the single factor market model Ri,t = α i β i Rmt ε it , where R i,tis the return of firm i and Rmt is the market return. In this study, regression analysis i s performed using 30 day ( -2 days to -31 days) return before the event date.


In this study, the following regression equation analyzes the effects of the SOA and corporate characteristics on firm value Because Black and Scholes (1974) point out that testing the effects of dividend policy on stock prices is the best method to examine firm valuation , either short term measures of stock price or risk adjusted returns have been used, while dividend smoothing policy holds for a long term period. Thus, Tobin’s Q, the ratio of a physical asset’s market value to its replacement va lue, is used to proxy firm value in this study To control for year and industry effects, year (YR) and industry (IND) variables are used.

Q = α1+ β1 SOA + β2 CashD + β3 ROA + β4 INV + β5 LEV + β6 SIZE + β7 CF + β8 A T + IND + YR + εi ............(4)

SOA represents dividend smoothness of Lintner’s (1956) partial adjustment model. The lower (higher) the SOA, the more (less) smoothed the dividends. SOA is measured using dividend per share (DPS) and earnings per share (EPS), as used by Fama & Babiak (1968), Fama (1974), Leary & Michaely (2011), Michaely & Roberts (2012), and Rhee and Park (2018). Lintner’s model measures the difference in firm dividend, obtained by applying the adjustment factor (ci) to the difference between the target dividend (D*it) and the previous dividend (Di,t−1). The target dividend is expressed as the target dividend propensity (ri) and the earning of the firm (D*it = riEit ). When applied to Eq. (5), it is expressed as Eq. (6), where β1 = ciri , β2 = −ci, respectively, and SOA is ci, i.e., −β2.

ΔDi,t = Di,t−Di,t−1 = αi + ci(D*it−Di,t−1) + εit ..............(5)

ΔDi,t = αi+ β1Eit + β2Di,t−1 + εit .................(6)

CashD represents the cash dividend amount and is measured as a function of the natural log of a firm’s cash dividend amount. To measure a firm’s profitability, return on assets (ROA) is used as the ratio of net income divided by total assets. Investment (INV) is the capital expenditure divided by total assets, and leverage (LEV) is the long term debt divided by total assets. Firm size (SIZE) is measured as the log of total assets. A firm’s cash flow ( CF) is measured as earnings before interest and taxes, plus depreciation less taxes, and normalized with total assets. A firm’s asset tangibility (AT ) is measured as the ratio of property, plant, and equipment divided by total assets.


This study used cash dividend data and financial statement data from companies listed on the Korea Exchange Market from 2000 to 2015. Financial statement data were extracted from TS 2000 and stock price data were collected using KIS VALUE. T he sample should have a minimum of 8 year s of DPS and EPS data for the period 2000 2015 to measure SOA. Zero DPS , dividend omissions, f inancial institutions and public institutions were excluded from the sample. For a legitimate analysis, firms are required to pay dividends for both the pre and post crisis period s . Of the total 742 firm s, the 10 8 companies that met the selection criteria were selected as the final research sample. Because t he financial crisis began in 2007 with the subprime mortgage shock in the U.S. and was moderated by the Fina ncial Stability Plan of the US Department of the Treasury in February 2009, the years from 2007 to 2009 are not included in the analysis. The 2001 2006 period was classified as pre crisis, and the 2010 2015 period was classified as post crisis.

Table 1 shows the summary statistics pre and post crisis. The post crisis SOA is 0.600 , which is about 5 % higher than the pre crisis level of 0.498. In line with Rhee and Park (2018), t his suggest s that firms avoid long term stable dividends in the aft ermath of a financial crisis. Post crisis ROA, INV, and CF are 0.039, 0.038, and 0.046, respectively, which are about 26.4%, 32.1%, and 28.1% higher than the pre crisis levels of 0.053, 0.056, and 0.064, respectively. Post crisis LEV, SIZE, and AT are 0.02 6, 13.060, and 0.433, respectively, which are about 36.8%, 2.1%, and 11.9% higher than the pre crisis levels of 0.019, 12.794, and 0.387, respectively. The results indicate that firms' profitability, investment level, and cash flow are de creased compared with those of pre crisis, while leverage, firm size and asset tangibility are in creased.

Table 1 Summary Statistics
  Pre- crisis Avg 0.498 8.020 0.053 0.056 0.019 12.794 0.064 0.387
Stdev 0.237 1.648 0.046 0.077 0.036 1.372 0.067 0.150
Max 0.991 13.628 0.246 0.314 0.249 17.873 0.338 0.860
Min 0.011 4.174 -0.113 -0.371 0.000 10.417 -0.117 0.062
N 329 329 329 329 329 329 329 329
  Post- crisis Avg 0.600 8.004 0.039 0.038 0.026 13.060 0.046 0.433
Stdev 0.246 1.591 0.058 0.156 0.046 1.320 0.058 0.195
Max 0.999 14.887 0.943 0.405 0.244 18.945 0.294 0.993
Min 0.016 1.099 -0.102 -1.895 0.000 10.570 -0.147 0.041
N 444 444 444 444 444 444 444 444
t-stat -5.779 0.133 3.579 2.110 -2.189 -2.729 3.775 -3.780

Table 2 provides the Pearson’s correlations of variables. Firms’ SOA has a negative and significant correlation with CashD, SIZE, and AT. CashD has a positive and significant correlation with ROA, INV, SIZE, CF, and AT, while a negative and significant correlation with LEV. ROA has a positive and significant correlation with SIZE and CF, while a negative and significant correlation with LEV. INV has a positive and significant correlation with LEV, SIZE, and CF. LEV has a positive and significant correlation with AT, while a negative and significant correlation with CF. SIZE has a positive and significant correlation with CF and AT.

Table2 Pearson's Correalatons     
SOA 1              
CashD -0.107
ROA -0.034 0.393 1          
  (0.350) (0.000)**  
INV 0.048 0.167 0.030 1        
  (0.180) (0.000)** (0.404)  
LEV 0.045 -0.122 -0.213 0.074 1      
  (0.216) (0.001)** (0.000)** (0.040)*  
SIZE -0.083 0.824 0.207 0.149 -0.008 1    
  (0.021)* (0.000)** (0.000)** (0.000)** (0.824)  
CF -0.015 0.404 0.422 0.136 -0.172 0.248 1  
  (0.681) (0.000)** (0.000)** (0.000)** (0.000)** (0.000)**  
AT -0.195 0.305 0.061 -0.032 0.076 0.420 0.002 1
  (0.000)** (0.000)** (0.088) (0.369) (0.036)* (0.000)** (0.967)  

Results and Discussion

Table 3 show's the pre and post crisis market response to dividend announcements using Eqs. (1)–(3). Pre crisis CARs of MktAdj, MeanAdj, and MktM are 0.0057, 0.0067, and 0.0059, respectively, and post crisis CARs are 0.0055, 0.0034, and 0.0040, respectively. Although each of the CAR models shows a positive market response for both pre and post crisis, the level of market response fell by 0.02%pt. in MktAdj , by 0.33% in MeanAdj , and by 0.19%pt. in MktM compared to pre crisis. The results indicate that post crisis, the market still reacts positively but less for dividend smoothing announcements . The results mean that market expectations for dividend smoothing have decreased , and support the hypothesis that the market response will be lowered than before the financial crisis.

Table 3 Pre- and Post-Crisis Cars
CAR MktAdj MeanAdj MktM
  Pre-crisis Avg 0.0057 0.0067 0.0059
Stdev 0.0506 0.0566 0.0541
Max 0.305 0.247 0.408
Min -0.152 -0.196 -0.197
N 329 329 326
  Post-crisis Avg 0.0055 0.0034 0.0040
Stdev 0.0469 0.0474 0.0490
Max 0.443 0.331 0.309
Min -0.153 -0.177 -0.279
N 444 444 443
t-stat 0.070 0.873 0.506

A market response comparative analysis for firms that increase or decrease SOA was conducted. The results are presented in Table 4. For firms that decrease the SOA, pre crisis CARs of MktAdj, MeanAdj, and MktM are 0.0022, 0.0032, and 0.0002, respectively, and post crisis CARs are 0.0070, 0.0043, and 0.0066, respectively. All three models show increased market response by 0.48%pt., 0.11%pt., and 0.64%pt., respectively, compared to the pre crisis market response. On the other hand, for firms that increase the SOA, pre crisis CARs of MktAdj, MeanAdj, and MktM are 0.0081, 0.0091, and 0.0097, respectively, and post crisis CARs are 0.0046, 0.0028, and 0.0025, respectively. All three models show decreased market respon se by 0.3%pt., 0.63%pt., and 0.72%pt., respectively, compared to the pre crisis period. Contrary to the assertion of Larkin et al. ( that dividend smoothing has little effect on stock prices, the results indicate that the market prefers more smoothed dividends and penalizes firms with volatile dividend s, s upporting the hypotheses .

Table 4 Pre- and Post-Crisis Cars According to Soa Increase and Decrease
CAR MktAdj MeanAdj MktM
    SOA Decrease   Pre-crisis Avg 0.0022 0.0032 0.0002
Stdev 0.0467 0.0543 0.0491
Max 0.1441 0.1671 0.1197
Min -0.1279 -0.1701 -0.1579
N 133 133 132
  Post-crisis Avg 0.0070 0.0043 0.0066
Stdev 0.0582 0.0552 0.0536
Max 0.4428 0.3307 0.3085
Min -0.1526 -0.1769 -0.1569
N 163 163 162
t-stat -0.7882 -0.1735 -1.0656
    SOA Increase   Pre-crisis Avg 0.0081 0.0091 0.0097
Stdev 0.0531 0.0581 0.0570
Max 0.3048 0.2468 0.4078
Min -0.1517 -0.1960 -0.1974
N 196 196 194
  Post-crisis Avg 0.0046 0.0028 0.0025
Stdev 0.0390 0.0424 0.0461
Max 0.1685 0.1782 0.1655
Min -0.1296 -0.1619 -0.2794
N 281 281 281
t-stat 0.7916 1.2967 1.4781

Table 5 presents the results of the effect of the SOA and firm characteristics on firm values using Eq. (4). In this regression, industry (IND) and year (YR) effects are controlled. Each regression explains 4.2% to 33.3% of the cross sectional variations in corporate values. The first two columns represent the regression results of firms that decr ease the SOA, and the last two columns represent the results of firms that increase the SOA . The specification in columns (1) and (3) include only the main variable to test the sensitivity of the firm value to the SOA. In other columns, variables for firm characteristics are included.

Table 5 Cross-Sectional Regression Analysis According to Soa Increase or  Decrease
Dep: Q SOA Decrease SOA Increase
(1) (2) (3) (4)
Intercept -48.30 (-2.507) -61.058(-3.409) -33.592(-3.874) -53.056(-6.969)
SOA -0.573(-3.609)*** -0.273(-1.844)* -0.045(-0.596) -0.098(-1.516)
CashD   0.181(3.969)***   0.087(5.343)***
ROA   -0.564(-1.019)   2.698(5.877)***
INV   0.175(0.769)   0.074(0.562)
LEV   -1.688(-1.930)*   0.346 (1.075)
SIZE   -0.126(-2.443)**   -0.045(-2.472)*
CF   1.808(2.746)***   0.524(2.055)**
AT   0.853(3.837)***   -0.048(-0.507)
Industry, Year Y Y Y Y
R_sq 0.125 0.333 0.042 0.327

In regressions (1) and (2), the coefficients of SOA are -0.573 and 0.273, respectively, and significant at the 1% and 10% levels, while in regressions (3) and (4), the coefficients of SOA are negative and insignificant. The results suggest that the speed of adjustment plays a critical factor in determining the value of firms that payout dividends in a more smooth manner while it is less critical for firms that less smooth ing dividends. Thus the results support the hypothesis.

In regression (2), the coefficients on CashD CF , and AT are positive and significant at the 1% levels and the coefficients on LEV and SIZE a re negative and significant at the 10% and 5%, respectively , imply ing that cash dividend amount , cash flow , and asset tangibility are positively correlated with firm values while leverage and firm size are negatively correlated. In regression (4), CashD ROA, and CF have positive and significant coefficients, and SIZE has a negative and significant coefficient, suggesting that cash dividend amount , and cash flow positively affect firm values, while firm size negatively affects firm values. The main dif ference s in the firm characteristics of the effect on firm valu ation of more and less dividend smooth ing firms are low leverage and high asset tangibility for more smoothing firms and high profitability for less smoothing firms.

Firstly, the post crisis SOA is higher than before the crisis and the level of market response has fallen. Second, the coefficient of SOA o n firm values for SOA decrease firms is negative and significant. The results suggest that conveying information thro ugh dividend smoothing has weakened since the financial crisis, and the market seems to have responded to it, and that dividend smoothing affects the value of firms with more dividend smooth ing Dividend payments of a compan y that follows dividend smoothin g are carried out regardless of company performance. The financial crisis affected the company's performance and cash generation, and there must have been difficulties in smoothing dividends. Therefore, it would be difficult for the compan y to continue wit h the existing dividend smoothing policy, and it would prefer a dividend policy that pays according to the company's performance. The characteristics of firms that affect the firm value s differed according to the degree of dividend smoothing. This would be an interesting extension of this study for future research.


This study incorporates two existing hypotheses on firms' dividend smoothing propensity and analyzes the impact on firm valuation. Consistent with the hypothesis, the post crisis SOA is 0.600, which is 20.5% higher than the pre crisis SOA of 0.498. The result suggests that, after the financial crisis, the degree of dividend smoothing decreased due to the financial market volatility.

In order to measure the market response to dividend announcement s , the market adjusted mode l the mean adjusted model, and the market model are used . Each CAR models show positive but decrease d market responses compared to pre crisis levels In addition, when firms payout dividends in a more smooth manner , the market shows a positive response of 0.11%pt. to 0.64%pt. compared to the pre crisis level s This result suggests that the market still favor s more smooth ing dividends but to a lesser extent than it was before the crisis.

For SOA decrease firms, the coefficient signs of SOA, LEV, and SIZE are negative and significant, and those of CashD, CF, and AT are positive and significant. The results suggest th at firm values with more smoothing dividend policy are affected by a more s mooth payout manner , less leverage, small firm size, more cash dividends, high cash flow, and high asset tangibility. For SOA increase firms, the coefficient signs of CashD , ROA, and CF are positive and significant, and that of SIZE is negative and significant. The results also suggest that firm values with less smoothed dividends are affected by more cash dividends, profitability, high cash flow, and small firm size. In sum , in Korea, the speed of adjustment, leverage, asset tangibility, and profitability affect firm value i n different ways according to the degree of dividend smoothing, but c ash dividend amount , firm size , and cash flow a ffect in the same way regardless of d ividend smoothing



  1. Aharony, J., & I. Swary (1980). Quarterly dividend and earnings announcements and stockholder returns: an empirical analysis. Journal of Finance, 35(1), 1–12.
  2. Alli, K., Q. Khan & G. Ramirez (1993). Determinants of corporate dividend policy: A factorial analysis.
  3. Financial Review, 28(4), 523–547.
  4. Berk, J. & P. DeMarzo (2013) Corporate Finance, Boston, MA: Prentice Hall.
  5. Bernhardt, D., A. Douglas & F. Robertson (2005). Testing dividend signaling models. Journal of Empirical Finance, 12(1), 77–98.
  6. Bhattacharya, S. (1979). Imperfect information, dividend policy, and ‘The bird in the hand’ fallacy. Bell Journal of Economics, 10(1), 259–270.
  7. Black, F. & M. Scholes (1974). The effect of dividend yield and dividend policy on common stock prices and returns. Journal of Financial Economics, 1, 1–22.
  8. Brav, A., J.R. Graham, C.R. Harvey & R. Michaely (2005). Payout policy in the 21st century. Journal of Financial Economics, 77(3), 483–527.
  9. Calvo, G. & E.G. Mendoza (2000). Rational contagion and the globalization of securities markets. Journal of International Economics, 51(1), 79–113.
  10. DeAngelo, H. & L. DeAngelo (2007). Capital structure, payout policy, and financial flexibility. Working paper, University of Southern California.
  11. DeMarzo, P. & Y. Sannikov (2016). Learning, termination, and payout policy in dynamic incentive contracts.
  12. Working paper, Stanford University.
  13. Fama, E. (1974). The empirical relationship between the dividend and the investment decisions of firms.
  14. American Economics Review, 64(3), 304–318.
  15. Fama, E. & H. Babiak (1968). Dividend policy: An empirical analysis. Journal of American Statistical Association, 63(324), 1132–1161.
  16. Guttman, I., O. Kadan & E. Kandel (2010). Dividend stickiness and strategic pooling. Review of Financial Studies, 23(10), 4455–4495.
  17. Kim, I.J. (2012). The global financial crisis and the challenges of the Korean economy. Seoul Journal of Economics, 25(3), 339–366.
  18. Lambrecht, B.M. & S.C. Myers (2012). A Lintner model of payout and managerial rents. Journal of Finance,
  19. 67(5), 1761–1810.
  20. Larkin, Y., M. Leary & R. Michaely (2017). Do investors value dividend smoothing stocks differently?
  21. Management Science, 63(12), 4114–4136.
  22. Leary, M. & R. Michaely (2011). Determinants of dividend smoothing: Empirical evidence. Review of Financial Studies, 24(10), 3197–3249.
  23. Lintner, J. (1956). Distribution of incomes of corporations among dividends, retained earnings, and taxes.
  24. American Economics Review, 46(2), 97–113.
  26. Mahmudi, H. & M. Pavlin (2013). Corporate payout policy, cash savings, and the cost of consistency: Evidence from a structural estimation. Financial Management, 42(4), 843–874.
  27. Michaely, R. & M. Roberts (2012). Corporate dividend policies: Lessons from private firms. Review of Financial Studies, 25(3), 711–746.
  28. Pettit, R. (1972). Dividend announcements, security performance, and capital market efficiency. Journal of Finance, 27(5), 993–1001.
  29. Rhee, K.J. & K.S. Park (2018). Changes in dividend smoothing after the financial crisis. Economics Letters, 172, 37–39.
  30. Ross, S. (1977). The determination of financial structure: The incentive-signalling approach. Bell Journal of Economics, 8(1), 23–40.
  31. Stiglitz, J.E. (2010). Lessons from the global financial crisis of 2008. Seoul Journal of Economics, 23(3), 321– 339.
Get the App