Dear Tax Expenditure Pre-conferencers:
I am looking forward to seeing you on Thursday morning for our
pre-conference meeting. I want to raise several issues that I hope you
will think about between now and then, and to ask you to be ready to
discuss several things at our meeting.
1. Distribution Tables. There seems to have been general agreement
that we can all use similar income cutoffs to make our tables
consistent, and that we can try to do these tables for a single
benchmark year such as 2006. However, we should talk about the
appropriate income concept and the issues that this will raise for
working with information from different data sets. In particular, if we
try to use AGI as the income concept for ranking, we run into the
problem of constructing AGI from the various surveys that may be part of
the various studies. Could you please be ready to talk about (a) all of
the household data sets that you plan to use in your study; (b) how
closely those data sets can be used to construct a measure of AGI? In
some cases, such as the SCF, there is already a programmed interface for
trying to do just this. Other data sets may require more work to get to
this point.
2. Matching Aggregates. One issue that will arise in all of our
studies concerns the degree of agreement between whatever micro data
sets we are analyzing and aggregate data such as NIPA information on
housing or health insurance, or aggregate data used by JCT and OTA in
making their tax expenditure estimates. I hope that for each study
where there is a natural aggregate "benchmark" against which the micro
data can be compared, that you will be able to discuss how the micro
data compare with the benchmark. We may want to discuss what the
appropriate benchmarks are for
3. Efficiency Analysis. One of the key issues we should discuss is how
to do the efficiency analysis associated with each of the tax
expenditures we are studying. This should lead us to a discussion of
parameter values such as demand elasticities that are going to be
central for the analysis, and to which estimates of these parameters are
going to be invoked in the analysis. For some papers this analysis is
probably straightforward, while for others it is more complex. Could
you please think about the issue of efficiency comparison between now
and Thursday morning?
Thank you for considering these issues. We will have powerpoint
capacity in our pre-conference room on Thursday morning, and I'll plan
to allow about 15 minutes for each project team to talk about their
work. We will use the rest of the time for more general discussion. I
look forward to seeing you in three days.
jim
Received on Mon Nov 05 2007 - 19:18:26 EST
