to 2001 which is outside of our sample period[21]. We also eliminate an additional 34
companies that made the freeze decision in conjunction with merger and acquisition
activity, as part of a subsidiary spin-off activity, simultaneously with a bankruptcy
filing, or with respect to non-US plans only. Consequently, our final base sample
consists of 147 freeze firms.
The remaining 1,111 DB firms form the population fromwhich we derive our matched
sample of 147 non-freeze firms. We randomly choose matching firms based on industry
and size. First, we match non-freeze firms with freeze firms based on two-digit SIC codes.
Then we pair freeze and non-freeze firms within two-digit SIC codes based on their total
assets to obtain our matched-pair sample. Once we pair a non-freeze firm with a freeze
firm, we remove that non-freeze firmfromthe population of eligiblematching firms.
Financial statement footnotes and COMPUSTAT serve as the primary sources of
data in our study. We also collect CEO tenure data from EXECUCOMP and, in some
cases, directly from company 10-K or proxy statement filings.We collect DB plan union
status data from Form 5,500 filings available on www.FreeERISA.com. Finally, we
eliminate firm-year observations from our base sample of 147 freeze firms and 147 nonfreeze
firms in the context of specific empirical tests due to missing data[22].
We present a distribution of firms announcing hard and soft/partial freeze decisions
over the sample period in panel A of Table II[23]. The decision is, for the most part,
evenly split between the different freeze types, though the incidence of soft/partial
freezes increases in 2006. It is apparent that there is a high concentration of freeze
announcements generally in the latter half of the period (2004-2006), which accounts
for about two-thirds of our sample firms. This is a period that follows the confluence of
various factors affecting DB plans – significant weakness in stock market returns
(which reduces the value of DB plan assets), low interest rates (which increases the
value of PBO), and proposals for extensive accounting and regulatory changes
affecting DB plans that were likely to be enacted.
Table II, panel B classifies sample firms by industry and freeze type. It is interesting
to note that the virtually identical split between hard and soft/partial freezes remains
consistent across various industries as well. There is also no significant industry
concentration among our sample firms; chemical firms (13.61 percent of our sample)
make up the largest industry group. We do, however, find a preponderance of
manufacturing firms in our freeze sample.
As a preliminary step in our investigation of the pension freeze decision, we examine
descriptive characteristics of freeze firms and non-freeze firms. Table III provides means,
standard deviations, medians, and quartiles for key variables used in our study, along
with univariate tests for differences in means and medians between freeze firms and
non-freeze firms.
Freeze firms have significantly poorer funded positions, as indicated by lower net
pension assets (NETPENASSET), relative to non-freeze firms. This is mainly due to
the significantly higher projected benefit obligation (PBO) of freeze firms. We also
decompose the funded status into the amount recognized on the balance sheet prior to
SFAS 158 (BS_RECOG) and the potential balance sheet impact of SFAS 158
(SFAS158_EFFECT). Interestingly, BS_RECOG for freeze firms is significantly
smaller and characterized by negative table values, whereas for non-freeze firms it is
characterized by positive table values. Also, freeze firms have a significantly larger
SFAS158_EFFECT. These univariate results suggest a potentially larger impact of the
anticipated new pension standard on freeze firms. This is consistent with freeze firms
responding to the likelihood of being required to recognize currently disclosed DB plan
to 2001 which is outside of our sample period[21]. We also eliminate an additional 34
companies that made the freeze decision in conjunction with merger and acquisition
activity, as part of a subsidiary spin-off activity, simultaneously with a bankruptcy
filing, or with respect to non-US plans only. Consequently, our final base sample
consists of 147 freeze firms.
The remaining 1,111 DB firms form the population fromwhich we derive our matched
sample of 147 non-freeze firms. We randomly choose matching firms based on industry
and size. First, we match non-freeze firms with freeze firms based on two-digit SIC codes.
Then we pair freeze and non-freeze firms within two-digit SIC codes based on their total
assets to obtain our matched-pair sample. Once we pair a non-freeze firm with a freeze
firm, we remove that non-freeze firmfromthe population of eligiblematching firms.
Financial statement footnotes and COMPUSTAT serve as the primary sources of
data in our study. We also collect CEO tenure data from EXECUCOMP and, in some
cases, directly from company 10-K or proxy statement filings.We collect DB plan union
status data from Form 5,500 filings available on www.FreeERISA.com. Finally, we
eliminate firm-year observations from our base sample of 147 freeze firms and 147 nonfreeze
firms in the context of specific empirical tests due to missing data[22].
We present a distribution of firms announcing hard and soft/partial freeze decisions
over the sample period in panel A of Table II[23]. The decision is, for the most part,
evenly split between the different freeze types, though the incidence of soft/partial
freezes increases in 2006. It is apparent that there is a high concentration of freeze
announcements generally in the latter half of the period (2004-2006), which accounts
for about two-thirds of our sample firms. This is a period that follows the confluence of
various factors affecting DB plans – significant weakness in stock market returns
(which reduces the value of DB plan assets), low interest rates (which increases the
value of PBO), and proposals for extensive accounting and regulatory changes
affecting DB plans that were likely to be enacted.
Table II, panel B classifies sample firms by industry and freeze type. It is interesting
to note that the virtually identical split between hard and soft/partial freezes remains
consistent across various industries as well. There is also no significant industry
concentration among our sample firms; chemical firms (13.61 percent of our sample)
make up the largest industry group. We do, however, find a preponderance of
manufacturing firms in our freeze sample.
As a preliminary step in our investigation of the pension freeze decision, we examine
descriptive characteristics of freeze firms and non-freeze firms. Table III provides means,
standard deviations, medians, and quartiles for key variables used in our study, along
with univariate tests for differences in means and medians between freeze firms and
non-freeze firms.
Freeze firms have significantly poorer funded positions, as indicated by lower net
pension assets (NETPENASSET), relative to non-freeze firms. This is mainly due to
the significantly higher projected benefit obligation (PBO) of freeze firms. We also
decompose the funded status into the amount recognized on the balance sheet prior to
SFAS 158 (BS_RECOG) and the potential balance sheet impact of SFAS 158
(SFAS158_EFFECT). Interestingly, BS_RECOG for freeze firms is significantly
smaller and characterized by negative table values, whereas for non-freeze firms it is
characterized by positive table values. Also, freeze firms have a significantly larger
SFAS158_EFFECT. These univariate results suggest a potentially larger impact of the
anticipated new pension standard on freeze firms. This is consistent with freeze firms
responding to the likelihood of being required to recognize currently disclosed DB plan
การแปล กรุณารอสักครู่..