Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
PhD em Economia
Universidade de Braslia
otavio@unb.br
RESUMO
O objetivo do trabalho identificar a dinmica lead-lag da relao entre lucro lquido e retorno
acionrio das empresas brasileiras de capital aberto. Testaram-se trs hipteses: a relao lucroretorno das empresas brasileiras de capital aberto dinmica distribuda ao longo do tempo; o
mercado se antecipa formao do lucro lquido do exerccio precificando as aes ao longo do
exerccio, de modo que os retornos lideram o lucro; o mercado se antecipa divulgao do lucro e
continua ajustando a precificao das aes no perodo entre o trmino do exerccio e a divulgao do
lucro lquido. A metodologia adotou a regresso reversa e dados em painel. A partir de janeiro do ano
corrente os retornos acionrios j refletem as informaes a respeito do lucro do prprio exerccio
corrente, corroborando que a relao lucros-retornos distribuda no tempo. Os retornos associados
aos meses de agosto, setembro, outubro e dezembro, entretanto, no se mostraram significantes. Isso
se deve, possivelmente, ao fato de as informaes transmitidas pelas empresas nesses meses no
terem sido consideradas suficientemente relevantes para alterar os preos das aes. Os resultados
evidenciaram que a relao lucro-retorno dinmica e o mercado se antecipa formao do lucro
precificando as aes ao longo do exerccio, todavia continua as ajustando aps o trmino do
exerccio.
Palavras-chave: Lucro contbil. Retornos acionrios. Relao lucros-retornos. Mercado acionrio.
Empresas brasileiras.
The lead lag dinamic between accounting earnings and stock returns
ABSTRACT
This paper is aimed at empirically identifying the lead-lag dynamics of the relationships between net
earnings and stock returns of publicly listed Brazilian companies. Three research hypotheses were
tested: the earnings-returns relationship is dynamic distributed through time; the stock market
anticipates the formation of the fiscal years net earnings, pricing stocks throughout the year, so that
returns lead earnings; the stock market anticipates the disclosure of net earnings and keep adjusting
the stock pricing during the period from the end of the fiscal year and the disclosure of net earnings.
The methodology is based on the reverse regressions between earnings and returns, and on
econometric panel data. The results show that the relationship between net earnings and stock returns
of Brazilian listed companies is dynamic and the stock market anticipates the formation of net earnings,
pricing stocks throughout the fiscal year until the disclosure of net earnings. The estimation by MQ2E in
panel data shows that from January of the current year, stock returns already reflect information
concerning the earnings of the year end, which corroborates that the earnings-return relationship is
distributed in time. However, stock returns associated to the months of August, September, October,
and December were not found to be significant. This is possibly due to the fact that the information
transmitted in these months was not considered sufficiently relevant by the market to change prices.
Enf.: Ref. Cont.
UEM - Paran
v. 34
n. 1
p. 103-121
104
The results show that the relationship between net earnings and stock returns is dynamic and the stock
market anticipates the formation of net earnings, pricing stocks throughout the fiscal year until the
disclosure of net earnings, nevertheless the market keeps adjusting stock prices after the end of the
fiscal year.
Keywords: Net earnings. Stock returns. Earnings-returns relationship. Stock market. Brazilian
companies.
INTRODUO
UEM - Paran
v. 34
p. 103-121
105
UEM - Paran
v. 34
p. 103-121
106
REFERENCIAL TERICO
UEM - Paran
v. 34
p. 103-121
UEM - Paran
v. 34
107
METODOLOGIA
p. 103-121
108
(2)
na qual:
Xit = LLit/VMit-1.
UEM - Paran
v. 34
p. 103-121
UEM - Paran
v. 34
109
p. 103-121
110
a)
b)
b)
UEM - Paran
v. 34
simultaneidade
p. 103-121
RESULTADOS E ANLISES
111
Estatstica
-259,6740
-970,1610
380,5500
308,6830
Prob. *
0,0000
0,0000
0,0000
0,0000
Cross-Sections
75
75
75
75
Obs.
581
581
581
600
Amostra: 2001 a 2009 - Variveis exgenas: Efeitos individuais - Seleo automtica do mximo de lags. *As probabilidades
para os testes de Fisher so computadas utilizando distribuio qui-quadrado assinttica. Todos os outros testes presumem
normalidade assinttica.
UEM - Paran
v. 34
padres
e
covarincia
robusta
n. 1
p. 103-121
112
UEM - Paran
v. 34
n. 1
Qui-Quadrado
15
p. 103-121
Probabilidade
0,0101
113
Coeficiente
-0,0598
0,1573
0,3340
0,1531
0,0557
0,1575
0,1934
0,3442
0,1124
-0,0915
0,1191
0,1773
0,1589
0,0926
0,2456
0,1206
0,5567
0,4892
8,2529
1,8559
Erro Padro
0,0075
0,0282
0,0885
0,0656
0,0226
0,0483
0,0333
0,0445
0,0524
0,0394
0,0818
0,0495
0,0753
0,0371
0,0693
0,0479
Estatstica t
-7,9882
5,5767
3,7722
2,3337
2,4697
3,2594
5,8119
7,7274
2,1447
-2,3220
1,4557
3,5787
2,1117
2,4935
3,5448
2,5197
Prob.
0,0000
0,0000
0,0002
0,0199
0,0138
0,0012
0,0000
0,0000
0,0324
0,0206
0,1460
0,0004
0,0351
0,0129
0,0004
0,0120
Obs.: Xit = LLit/VMit-1. LLit = lucro lquido do exerccio da i-sima empresa na data t. VMit-1 =
valor de mercado da i-sima empresa na data t-1. Ri = retorno acionrio mensal da i-sima
empresa na data . = t-m, ..., t-1, t, t+1, ..., t+n. m = nmero de lags. n = nmero de leads. i =
intercepto da equao (3). i = coeficiente angular da equao (3). ut = termo aleatrio, no qual
ut ~ N(0, 2) .
Coeficiente
-0,0645
0,1399
0,3047
0,1423
0,0546
0,1601
0,1825
0,3529
0,0844
-0,0758
0,1128
0,1879
0,2745
0,0879
0,2341
0,1176
0,5489
0,4802
8,0869
1,8914
Erro Padro
0,0122
0,0309
0,0796
0,0673
0,0227
0,0499
0,0324
0,0670
0,0538
0,0458
0,0952
0,0627
0,1710
0,0404
0,0618
0,0451
Estatstica t
-5,2662
4,5275
3,8290
2,1148
2,4014
3,2083
5,6348
5,2638
1,5688
-1,6551
1,1845
2,9981
1,6050
2,1757
3,7869
2,6072
Prob.
0,0000
0,0000
0,0001
0,0349
0,0166
0,0014
0,0000
0,0000
0,1172
0,0985
0,2367
0,0028
0,1090
0,0300
0,0002
0,0094
Obs.: Xit = LLit/VMit-1. LLit = lucro lquido do exerccio da i-sima empresa na data t. VMit-1 = valor de
mercado da i-sima empresa na data t-1. Ri = retorno acionrio mensal da i-sima empresa na data
. = t-m, ..., t-1, t, t+1, ..., t+n. m = nmero de lags. n = nmero de leads. i = intercepto da
equao (3). i = coeficiente angular da equao (3). ut = termo aleatrio, no qual ut ~ N(0, 2).
UEM - Paran
v. 34
painel
com
efeitos
fixos)
e
a
heteroscedasticidade dos resduos (utilizando
erros padres de White).
Os coeficientes so significantes em conjunto,
conforme o teste F (prob. 0,0000). So, tambm,
n. 1
p. 103-121
114
UEM - Paran
v. 34
p. 103-121
CONSIDERAES FINAIS
UEM - Paran
v. 34
115
p. 103-121
116
REFERNCIAS
BALL, R. SHIVAKUMAR, L. How much new
information is there in earnings? Journal of
Accounting Research, v. 46, n. 5, p. 975-1016,
dez. 2008.
BALL, R.; BROWN, P. An empirical evaluation of
accounting income numbers. Journal of
Accounting Research, v. 6, n. 6, p. 159-178,
outono, 1968.
BASU, S. The conservatism principle and the
asymmetric timeliness of earnings. Journal of
Accounting and Economics, v. 24, n. 1, p. 3-37,
dez. 1997.
BEAVER, W. H. Perspectives on Recent Capital
Market Research, The Accounting Review,
Enf.: Ref. Cont.
UEM - Paran
v. 34
p. 103-121
117
UEM - Paran
v. 34
p. 103-121
118
accounting
conservatism.
Journal
of
Accounting and Economics, v. 48, p. 132-150,
2009.
KORMENDI, R.; LIPE, R. Earnings innovations,
earnings persistence and stock returns. Journal
of Business, v. 60, n. 2, p. 323-345, jul. 1987.
KOTHARI, S. P. Price-earnings regressions in
the presence of prices leading earnings. Journal
of Accounting and Economics, v. 15, n. 2-3, p.
173-202, ago. 1992.
KOTHARI, S. P.; LEWELLEN, J.; WARNER, J. B.
Stock returns and aggregate earnings surprises,
and behavioral finance. Journal of Financial
Economics, v. 79, p. 537-568, 2006.
LEV, B. On the usefulness of earnings and
earnings research: lessons and directions from
two decades of empirical research. Journal of
Accounting Research, v. 27, p. 153-192,
suplemento, 1989.
LEV, B.; THIAGARAJAN, R. Fundamental
information analysis. Journal of accounting
Research, v. 31, n. 2, p. 190-215, outono, 1993.
LEV, B.; ZAROWIN, P. The boundaries of.
financial reporting and how to extend them.
Journal of accounting Research, v. 37, n. 2, p.
353-385, outono, 1999.
LIM, S. C.; PARK, T. The declining association
between earnings and returns: diminishing value
relevance of earnings or noisier markets?
Management Research Review, v. 34, n. 8, p.
947-961, 2011.
LIPE, R.; KORMEDI, R. Mean reversion in annual
earnings and its implications for security
valuation. Review of Quantitative Finance and
Accounting, v. 4, n. 1, p. 27-46, 1994.
UEM - Paran
v. 34
n. 1
p. 103-121
UEM - Paran
v. 34
119
n. 1
p. 103-121
120
Coeficiente
0,0534
0,1670
0,4090
0,1892
0,1567
0,2383
0,3063
0,3952
0,1826
-0,1115
0,1310
0,1616
0,2064
0,0891
0,2746
0,2535
0,2933
0,2772
18,2357
1,5369
Erro Padro
0,0124
0,0399
0,1050
0,0692
0,0286
0,0774
0,0486
0,0741
0,0634
0,0485
0,0955
0,0695
0,0605
0,0372
0,0645
0,0694
Estatstica t
4,3022
4,1847
3,8970
2,7329
5,4834
3,0798
6,3028
5,3337
2,8810
-2,2993
1,3712
2,3258
3,4151
2,3940
4,2597
3,6530
Prob.
0,0000
0,0000
0,0001
0,0064
0,0000
0,0022
0,0000
0,0000
0,0041
0,0218
0,1708
0,0203
0,0007
0,0169
0,0000
0,0003
Coeficiente
-0,0598
0,1573
0,3340
0,1531
0,0557
0,1575
0,1934
0,3442
0,1124
-0,0915
0,1191
0,1773
0,1589
0,0926
0,2456
0,1206
0,5567
0,4892
8,2529
1,8559
UEM - Paran
Erro Padro
0,0075
0,0282
0,0885
0,0656
0,0226
0,0483
0,0333
0,0445
0,0524
0,0394
0,0818
0,0495
0,0753
0,0371
0,0693
0,0479
v. 34
n. 1
Estatstica t
-7,9882
5,5767
3,7722
2,3337
2,4697
3,2594
5,8119
7,7274
2,1447
-2,3220
1,4557
3,5787
2,1117
2,4935
3,5448
2,5197
p. 103-121
Prob.
0,0000
0,0000
0,0002
0,0199
0,0138
0,0012
0,0000
0,0000
0,0324
0,0206
0,1460
0,0004
0,0351
0,0129
0,0004
0,0120
121
Coeficiente
-0.2446
0.4916
0.0101
-0.1816
0.2179
0.0740
-0.2989
0.4894
0.1662
0.0359
-0.1399
-0.1299
0.3019
0.1920
0.2688
0.1276
0.1091
0.2878
0.1515
0.5051
-0.3377
0.1348
-0.0711
-0.0170
3.5240
0.0524
0.2169
-0.1106
0.4691
-0.2493
-0.2824
0.1248
-0.0349
7.1475
1.9859
Erro Padro
0.2012
0.1709
0.1673
0.4497
0.1482
0.2545
0.2069
0.2659
0.2382
0.1692
0.3100
0.1342
0.3297
0.2012
0.2892
0.1913
0.1122
0.1093
0.1731
0.2281
0.4623
0.2158
0.5374
0.5908
3.9647
0.1605
0.1097
0.3000
0.4712
0.3168
0.6524
Estatstica t
-1.2153
2.8759
0.0605
-0.4037
1.4702
0.2908
-1.4445
1.8407
0.6979
0.2120
-0.4512
-0.9681
0.9155
0.9540
0.9296
0.6671
0.9721
2.6333
0.8751
2.2144
-0.7304
0.6246
-0.1322
-0.0288
0.8889
0.3268
1.9777
-0.3688
0.9954
-0.7869
-0.4328
Prob.
0.2248
0.0042
0.9518
0.6865
0.1421
0.7713
0.1491
0.0662
0.4856
0.8322
0.6520
0.3334
0.3603
0.3405
0.3530
0.5050
0.3314
0.0087
0.3819
0.0272
0.4655
0.5325
0.8949
0.9770
0.3745
0.7440
0.0484
0.7124
0.3199
0.4316
0.6653
Varivel dependente: X. Amostra: 2001 2009. Observaes includas: 9. Cross-sections includas: 75. Informaes pooled
(balanceadas) totais: 675. Lista de instrumentos: Rt-23, Rt-22, Rt-21, Rt-20, Rt-19, Rt-18, Rt-17, Rt-16, Rt-15, Rt-14, Rt-13, Rt-12, Rt-11, Rt-10, Rt9, Rt-8, Rt-7, Rt-6, Rt-5, Rt-4, Rt-3, Rt-2, Rt-1, Rt+1, Rt+2, Rt+3, Rt+4, Rt+5, Rt+6, Rt+7. Erro-padro e covarincia cross-section de White.
UEM - Paran
v. 34
n. 1
p. 103-121