Lagging a variable 1 or more years from data on daily stock prices to use in OLS regression

[Peter Ljungbeck]

I am trying to make an empirical study on the effect of dividend changes on future firm performance. I, therefore, want to regress dividend change on a measure of firm performance, in this case, it is Return on assets. I, however, need to run the regression using the return on assets in one and two years. My regression would like this:

reg roai2y divchange if divchange < -.20 & dummy == 1| divchange > .25 & divchange < 5 & dummy == 1

My problem is, therefore, how do I lag my return on assets variable, 1 or more years when I have multiple firms each with a different number of trading days in a year over 20 years. I use the dummy variable to indicate a change in dividend change. fyear is fiscal year and days is the observation day in the fiscal year. My data looks like this:

clear
input long datadate double fyear str88 conm double(cshoc prccd) float(divd divchange roai dummy) int days
15160 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82000 . .3636364 .009597253 . 112
15161 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82500 . .3636364 .009597253 . 113
15162 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82500 . .3636364 .009597253 . 114
15165 2001 "A.P. MOLLER - MAERSK A/S A" 372600 81500 . .3636364 .009597253 . 115
15166 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82500 . .3636364 .009597253 . 116
15167 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82500 . .3636364 .009597253 . 117
15168 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82500 . .3636364 .009597253 . 118
15169 2001 "A.P. MOLLER - MAERSK A/S A" 372600 81500 . .3636364 .009597253 . 119
15172 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82000 . .3636364 .009597253 . 120
15174 2001 "A.P. MOLLER - MAERSK A/S A" 372600 80000 . .3636364 .009597253 . 121
15175 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82000 . .3636364 .009597253 . 122
15176 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82500 . .3636364 .009597253 . 123
15180 2001 "A.P. MOLLER - MAERSK A/S A" 372600 81500 . .3636364 .009597253 . 124
15181 2001 "A.P. MOLLER - MAERSK A/S A" 372600 80000 . .3636364 .009597253 . 125
15182 2001 "A.P. MOLLER - MAERSK A/S A" 372600 79000 . .3636364 .009597253 . 126
15183 2001 "A.P. MOLLER - MAERSK A/S A" 372600 80000 . .3636364 .009597253 . 127
15186 2001 "A.P. MOLLER - MAERSK A/S A" 372600 80000 . .3636364 .009597253 . 128
15187 2001 "A.P. MOLLER - MAERSK A/S A" 372600 80500 . .3636364 .009597253 . 129
15188 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82000 . .3636364 .009597253 . 130
15189 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82000 . .3636364 .009597253 . 131
15190 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82000 . .3636364 .009597253 . 132
15193 2001 "A.P. MOLLER - MAERSK A/S A" 372600 80000 . .3636364 .009597253 . 133
15196 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82000 . .3636364 .009597253 . 134
15197 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82000 . .3636364 .009597253 . 135
15200 2001 "A.P. MOLLER - MAERSK A/S A" 372600 83000 . .3636364 .009597253 . 136
15201 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82500 . .3636364 .009597253 . 137
15202 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82500 . .3636364 .009597253 . 138
15207 2001 "A.P. MOLLER - MAERSK A/S A" 372600 80000 . .3636364 .009597253 . 139
15208 2001 "A.P. MOLLER - MAERSK A/S A" 372600 79500 . .3636364 .009597253 . 140
15214 2001 "A.P. MOLLER - MAERSK A/S A" 372600 82000 . .3636364 .009597253 . 141
15215 2001 "A.P. MOLLER - MAERSK A/S A" 372600 77500 . .3636364 .009597253 . 142
15216 2001 "A.P. MOLLER - MAERSK A/S A" 372600 74000 . .3636364 .009597253 . 143
15217 2001 "A.P. MOLLER - MAERSK A/S A" 372600 73200 . .3636364 .009597253 . 144
15218 2001 "A.P. MOLLER - MAERSK A/S A" 372600 75500 . .3636364 .009597253 . 145
15221 2001 "A.P. MOLLER - MAERSK A/S A" 372600 75000 . .3636364 .009597253 . 146
15222 2001 "A.P. MOLLER - MAERSK A/S A" 372600 73500 . .3636364 .009597253 . 147
15223 2001 "A.P. MOLLER - MAERSK A/S A" 372600 76000 . .3636364 .009597253 . 148
15224 2001 "A.P. MOLLER - MAERSK A/S A" 372600 76000 . .3636364 .009597253 . 149
15229 2001 "A.P. MOLLER - MAERSK A/S A" 372600 72000 . .3636364 .009597253 . 150
15231 2001 "A.P. MOLLER - MAERSK A/S A" 372600 72500 . .3636364 .009597253 . 151
15232 2001 "A.P. MOLLER - MAERSK A/S A" 372600 70000 . .3636364 .009597253 . 152
15235 2001 "A.P. MOLLER - MAERSK A/S A" 372600 67000 . .3636364 .009597253 . 153
15236 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65500 . .3636364 .009597253 . 154
15237 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65300 . .3636364 .009597253 . 155
15239 2001 "A.P. MOLLER - MAERSK A/S A" 372600 58500 . .3636364 .009597253 . 156
15242 2001 "A.P. MOLLER - MAERSK A/S A" 372600 61000 . .3636364 .009597253 . 157
15243 2001 "A.P. MOLLER - MAERSK A/S A" 372600 60000 . .3636364 .009597253 . 158
15244 2001 "A.P. MOLLER - MAERSK A/S A" 372600 59500 . .3636364 .009597253 . 159
15245 2001 "A.P. MOLLER - MAERSK A/S A" 372600 59500 . .3636364 .009597253 . 160
15246 2001 "A.P. MOLLER - MAERSK A/S A" 372600 58000 . .3636364 .009597253 . 161
15249 2001 "A.P. MOLLER - MAERSK A/S A" 372600 54000 . .3636364 .009597253 . 162
15250 2001 "A.P. MOLLER - MAERSK A/S A" 372600 55000 . .3636364 .009597253 . 163
15251 2001 "A.P. MOLLER - MAERSK A/S A" 372600 57000 . .3636364 .009597253 . 164
15252 2001 "A.P. MOLLER - MAERSK A/S A" 372600 61000 . .3636364 .009597253 . 165
15253 2001 "A.P. MOLLER - MAERSK A/S A" 372600 62500 . .3636364 .009597253 . 166
15256 2001 "A.P. MOLLER - MAERSK A/S A" 372600 60000 . .3636364 .009597253 . 167
15258 2001 "A.P. MOLLER - MAERSK A/S A" 372600 64500 . .3636364 .009597253 . 168
15259 2001 "A.P. MOLLER - MAERSK A/S A" 372600 66800 . .3636364 .009597253 . 169
15260 2001 "A.P. MOLLER - MAERSK A/S A" 372600 67000 . .3636364 .009597253 . 170
15263 2001 "A.P. MOLLER - MAERSK A/S A" 372600 64000 . .3636364 .009597253 . 171
15266 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65500 . .3636364 .009597253 . 172
15270 2001 "A.P. MOLLER - MAERSK A/S A" 372600 64000 . .3636364 .009597253 . 173
15271 2001 "A.P. MOLLER - MAERSK A/S A" 372600 63000 . .3636364 .009597253 . 174
15272 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65000 . .3636364 .009597253 . 175
15273 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65000 . .3636364 .009597253 . 176
15274 2001 "A.P. MOLLER - MAERSK A/S A" 372600 69500 . .3636364 .009597253 . 177
15278 2001 "A.P. MOLLER - MAERSK A/S A" 372600 66000 . .3636364 .009597253 . 178
15285 2001 "A.P. MOLLER - MAERSK A/S A" 372600 70900 . .3636364 .009597253 . 179
15286 2001 "A.P. MOLLER - MAERSK A/S A" 372600 67000 . .3636364 .009597253 . 180
15287 2001 "A.P. MOLLER - MAERSK A/S A" 372600 67000 . .3636364 .009597253 . 181
15291 2001 "A.P. MOLLER - MAERSK A/S A" 372600 62500 . .3636364 .009597253 . 182
15292 2001 "A.P. MOLLER - MAERSK A/S A" 372600 64700 . .3636364 .009597253 . 183
15293 2001 "A.P. MOLLER - MAERSK A/S A" 372600 67000 . .3636364 .009597253 . 184
15294 2001 "A.P. MOLLER - MAERSK A/S A" 372600 66500 . .3636364 .009597253 . 185
15295 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65000 . .3636364 .009597253 . 186
15298 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65000 . .3636364 .009597253 . 187
15299 2001 "A.P. MOLLER - MAERSK A/S A" 372600 66000 . .3636364 .009597253 . 188
15300 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65000 . .3636364 .009597253 . 189
15301 2001 "A.P. MOLLER - MAERSK A/S A" 372600 66000 . .3636364 .009597253 . 190
15305 2001 "A.P. MOLLER - MAERSK A/S A" 372600 64000 . .3636364 .009597253 . 191
15307 2001 "A.P. MOLLER - MAERSK A/S A" 372600 64000 . .3636364 .009597253 . 192
15308 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65000 . .3636364 .009597253 . 193
15309 2001 "A.P. MOLLER - MAERSK A/S A" 372600 67000 . .3636364 .009597253 . 194
15312 2001 "A.P. MOLLER - MAERSK A/S A" 372600 70000 . .3636364 .009597253 . 195
15313 2001 "A.P. MOLLER - MAERSK A/S A" 372600 68500 . .3636364 .009597253 . 196
15314 2001 "A.P. MOLLER - MAERSK A/S A" 372600 69900 . .3636364 .009597253 . 197
15315 2001 "A.P. MOLLER - MAERSK A/S A" 372600 69500 . .3636364 .009597253 . 198
15316 2001 "A.P. MOLLER - MAERSK A/S A" 372600 69000 . .3636364 .009597253 . 199
15319 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65000 . .3636364 .009597253 . 200
15320 2001 "A.P. MOLLER - MAERSK A/S A" 372600 67000 . .3636364 .009597253 . 201
15321 2001 "A.P. MOLLER - MAERSK A/S A" 372600 68000 . .3636364 .009597253 . 202
15322 2001 "A.P. MOLLER - MAERSK A/S A" 372600 68000 . .3636364 .009597253 . 203
15323 2001 "A.P. MOLLER - MAERSK A/S A" 372600 66000 . .3636364 .009597253 . 204
15326 2001 "A.P. MOLLER - MAERSK A/S A" 372600 64000 . .3636364 .009597253 . 205
15327 2001 "A.P. MOLLER - MAERSK A/S A" 372600 66000 . .3636364 .009597253 . 206
15328 2001 "A.P. MOLLER - MAERSK A/S A" 372600 64500 . .3636364 .009597253 . 207
15329 2001 "A.P. MOLLER - MAERSK A/S A" 372600 64900 . .3636364 .009597253 . 208
15330 2001 "A.P. MOLLER - MAERSK A/S A" 372600 65000 . .3636364 .009597253 . 209
15336 2001 "A.P. MOLLER - MAERSK A/S A" 372600 68000 . .3636364 .009597253 . 210
15337 2001 "A.P. MOLLER - MAERSK A/S A" 372600 69000 . .3636364 .009597253 . 211
end
format %d datadate

[Clyde Schechter]

Well, if your data example is reflective of your data as a whole, both the divchange variable and the roai variable are constant within a given fiscal year for the firm. So you don't really have daily data here. For the key variables in your model you have yearly data repeated many times. So it looks like you should reduce your data set to one observation per conm per fyear. Then you can create a numeric identifier for each conm, let's call it n_conm, and -xtset n_conm fyear- and then use Stata's ordinary lag operator (-help tsvarlist- if you aren't familiar with it.)

[Phil Bromiley]

I don't think it makes much sense to run a model of ROA or dividend change (either quarterly or annually) using daily data. If you use annual data, you're essentially entering 200 duplicate observations for each year. Since dividend change and ROA are reported either quarterly or annually, it would probably make sense to collapse your data to quarterly or annual. Once you do that, the lags become easy.

[Peter Ljungbeck]

Thank you for your responses.

I ended up using F365.roai but this ignores leap years. The reason I need daily prices is to calculate the beta value both to calculate an expected return variable so that an abnormal return variable can be calculated but also cause I am planning to examine the effects of a significant dividend change on a firm's systematic risk as measured by the beta.