Dummy variable regression interpretation

[Conrad Taylor]

Hello everyone, need some help
I created dumym varaibles for when my varaible dINDPRO (industry Production index) is in contraction
generate DUM1990 = 0
replace DUM1990 = 1 if Month>tm(1990m6) & Month<tm(1991m3))

I did this for 1999 dot come crash and 2008 recession.

regress ExcessRChevron ESRetS&P TermSpread dM1SUPPLY dINDPRO dCCREDIT dBMINUSA dInflation DUM1990 DUM2001 DUM2008

What does that output of the regression mean? Will the dummy capture the effects of the times when dINDPRO is in contraction or will it capture the effects when dINDPRO is in expansion?

Dummy Coefficient p value
Dum1990. .00224. 0.927
dum2001. .01289 0.592
dum2008. .015447 0.418

[Clyde Schechter]

Think of the months that are not singled out (value = 1) by any of the Dum1990, d7m2001, and dum2008 variables, as "normal times."
Then, for example, for Dum1990, the meaning is that, all else equal, the expected value of ExcessRChevron is 0.00224 lower for observations where Month is between July 1990 and February 1991 (inclusive on both ends) than during "normal times." This might also be referred to as the marginal effect of the 1990 event.

[Conrad Taylor]

so essentially the Dummies i have created are singling out the effects of the months i have specified?

[Clyde Schechter]

so essentially the Dummies i have created are singling out the effects of the months i have specified?

Loosely speaking, yes. But speaking loosely is hazardous in statistics. Your Dum1990 variable, for example, does not single out the effects of those months relative to any of the months covered by any of the other dum* variables. It "singles them out" only relative to the months not covered by any other dum* variable.