I am trying to estimate the effect of a lagged independent variable on wealth accumulation. Specifically, I am trying to estimate whether changes in the value of several asset/debt categories over 2007 to 2009 (value of equity, home value, etc.) explain future wealth growth. In other words, does a change in asset values over 2007-2009 explain the change in wealth from 2010 through 2019. I have a cohort-year level dataset (255 cohorts 5 panels/years 2007-2019 every 3 years). An example of the data is at the bottom of this post. I am wondering whether the below specification is appropriate for what I am trying to test.
I have tried the following specification:
Code:
reghdfe lnnetworth c.dlnliq0709##i.panel c.dlnequity0709##i.panel c.dlnfixedinc0709##i.panel c.dlncashli0709##i.panel c.dlnretqliq0709##i.panel c.dlnhouses0709##i.panel c.dlnbus0709##i.panel c.dlnedninst0709##i.panel c.dlnvehinst0709##i.panel c.dlnothinst0709##i.panel c.dlnccbal0709##i.panel c.dlnmrthel0709##i.panel [controls], a(cohort) vce(cluster cohort)
Array
My interpretation of the coefficient on dlnliq0709#2013 is that a one standard deviation increase in dlnliq0709 results in a ~4% increase in wealth in 2013 relative to 2007 wealth. Or should this be interpreted as a one sd increase in dlnliq0709 is associated with a ~4% greater increase in wealth relative to the effect of dlnliq0709 on 2007 wealth?
An alternative approach that I was planning on trying was to reshape (wide) into cohort-level data and simply regress the change in wealth from 2010 to 2019 on the asset/debt changes along with cohort-level controls (any obvious problems with this approach?).
Any guidance would be immensely appreciated.
Code:
* Example generated by -dataex-. For more info, type help dataex clear input float(cohort panel lnnetworth dlnliq0709 dlnequity0709 dlnfixedinc0709 dlncashli0709 dlnretqliq0709 dlnhouses0709 dlnbus0709 dlninstall0709 dlnccbal0709 dlnmrthel0709 dlnedninst0709 dlnvehinst0709 dlnothinst0709 dlnothloc0709 dlnodebt0709) 113 2007 12.356108 -.8117663 -.6679196 -.03997894 -2.414839 -.7627107 -.22390684 .09321466 .51196074 .7042804 -1.890627 0 -.04455568 .5565164 0 -.7152216 113 2010 12.369104 -.8117663 -.6679196 -.03997894 -2.414839 -.7627107 -.22390684 .09321466 .51196074 .7042804 -1.890627 0 -.04455568 .5565164 0 -.7152216 113 2013 12.067015 -.8117663 -.6679196 -.03997894 -2.414839 -.7627107 -.22390684 .09321466 .51196074 .7042804 -1.890627 0 -.04455568 .5565164 0 -.7152216 113 2016 12.37533 -.8117663 -.6679196 -.03997894 -2.414839 -.7627107 -.22390684 .09321466 .51196074 .7042804 -1.890627 0 -.04455568 .5565164 0 -.7152216 113 2019 12.415668 -.8117663 -.6679196 -.03997894 -2.414839 -.7627107 -.22390684 .09321466 .51196074 .7042804 -1.890627 0 -.04455568 .5565164 0 -.7152216 114 2007 12.860944 .1382894 .50214225 .065634914 .1201632 .55576324 .2047282 .0009744643 .1952882 -.09859852 .07706435 -.099766 .3135726 -.14225069 .21472567 -.3869707 114 2010 12.641363 .1382894 .50214225 .065634914 .1201632 .55576324 .2047282 .0009744643 .1952882 -.09859852 .07706435 -.099766 .3135726 -.14225069 .21472567 -.3869707 114 2013 12.510736 .1382894 .50214225 .065634914 .1201632 .55576324 .2047282 .0009744643 .1952882 -.09859852 .07706435 -.099766 .3135726 -.14225069 .21472567 -.3869707 114 2016 12.70246 .1382894 .50214225 .065634914 .1201632 .55576324 .2047282 .0009744643 .1952882 -.09859852 .07706435 -.099766 .3135726 -.14225069 .21472567 -.3869707 114 2019 12.407745 .1382894 .50214225 .065634914 .1201632 .55576324 .2047282 .0009744643 .1952882 -.09859852 .07706435 -.099766 .3135726 -.14225069 .21472567 -.3869707 115 2007 13.686165 -.4298446 -1.1421677 -.04004299 -.3491879 -1.2077113 -.03302917 -.4818779 -.185182 -.6227583 -.027859643 -.000841175 -.4847591 .3453323 -.11744729 -.12501265 115 2010 13.572906 -.4298446 -1.1421677 -.04004299 -.3491879 -1.2077113 -.03302917 -.4818779 -.185182 -.6227583 -.027859643 -.000841175 -.4847591 .3453323 -.11744729 -.12501265 115 2013 13.558372 -.4298446 -1.1421677 -.04004299 -.3491879 -1.2077113 -.03302917 -.4818779 -.185182 -.6227583 -.027859643 -.000841175 -.4847591 .3453323 -.11744729 -.12501265 115 2016 13.502767 -.4298446 -1.1421677 -.04004299 -.3491879 -1.2077113 -.03302917 -.4818779 -.185182 -.6227583 -.027859643 -.000841175 -.4847591 .3453323 -.11744729 -.12501265 115 2019 13.198543 -.4298446 -1.1421677 -.04004299 -.3491879 -1.2077113 -.03302917 -.4818779 -.185182 -.6227583 -.027859643 -.000841175 -.4847591 .3453323 -.11744729 -.12501265 116 2007 12.017576 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 116 2010 12.119977 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 116 2013 11.937418 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 116 2016 12.01959 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 116 2019 11.831182 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 117 2007 11.968186 .0031573025 0 0 1.8498133 0 -.09501783 0 3.2656076 1.6116536 -2.2984762 0 3.2656076 0 0 0 117 2010 10.896545 .0031573025 0 0 1.8498133 0 -.09501783 0 3.2656076 1.6116536 -2.2984762 0 3.2656076 0 0 0 117 2013 11.556764 .0031573025 0 0 1.8498133 0 -.09501783 0 3.2656076 1.6116536 -2.2984762 0 3.2656076 0 0 0 117 2016 12.021815 .0031573025 0 0 1.8498133 0 -.09501783 0 3.2656076 1.6116536 -2.2984762 0 3.2656076 0 0 0 117 2019 11.96965 .0031573025 0 0 1.8498133 0 -.09501783 0 3.2656076 1.6116536 -2.2984762 0 3.2656076 0 0 0 end
**DATA NOTE: the 2007 panel respondents were re-interviewed in 2009, allowing me to compute within-individual changes for the 2007 survey. These changes are then averaged at the cohort level and applied to each cohort for the entire dataset (i.e. not time-variant). Data from 2010-2019 are from later survey panels with a random selection of individuals which is why I am performing this analysis at the cohort level. Cohorts are defined by 3-year birth year, race, and level of education.
0 Response to Estimating effect of lagged changes on wealth accumulation
Post a Comment