Full text of "Asset accumulation in early married life"

See other formats

UNIVERSITY Of
ILLINOIS LIBRARY
AT URBANA-CHA^4PAlGt4
^ STACKS

e.24 5^-3

CENTRAL CIRCULATION BOOKSTACKS

The person charging this material is re-
sponsible for its renewal or its return to
the library from which it was borrowed
on or before the Latest Date stampd
below. You may be charged a minimum
fee of $75.00 for each lost book.

Theft, nwtllatloiv ond underlining of boolcs are reosons
for disciplinary o<t!on and may result In dismissal from
the University.
TO RENEW CAIL TELEPHONE CENTER, 333-8400

UNIVEIiSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN

J^L 2 5 1995

When renewing by phone, write new due date below
previous due date. L162

Digitized by the Internet Archive

in 2011 with funding from

University of Illinois Urbana-Champaign

http://www.archive.org/details/assetaccumulatio403leel

Faculty Working Papers

l,(oO

ASSET ACCUMULATION IN EARLY MARRIED LIFE*

-^^^

College of Commerce and Business Administration

University of Illinois at Urbana-Champaign

ASSET ACCUMULATION IN EARLY MARRIED LIFE*

79^^^^

Lucy Chao Lee and Robert Ferber

*This paper was prepared as part of the work under Grant
SOC74-23458 of the National Science Foundation.

FACULTY WORKING PAPERS
College of Commerce and Business Administration
University of Illinois at Urbana-Champaign

June 1, 1977

ASSET ACCUMULATION IN EARLY MARRIED LIFE
Lucy Chao Lee and Robert Ferber

#403

jfUJ<

MT m.T?'A '•! -irOA T?Ta?.A

Abstract
This study investigates the extent to which asset accumulation by
young married couples in early married life can be explained by financial,
other economic and attitudinal variables. It finds, among other things,
that at least for this sample, those who started out better off kept getting
still better off financially, and that going into debt was a major cor-
relate of later financial well being.

«'•<»

Asset Accumulation in Early Married Li f e

I . Introductior

What influences the accumulation of financial assets by young married
couples? Do couples that start out with more assets tend to maintain
this lead over time? Do couples that borrow tend to be better or worse
off financially than those who do not? To vvhat extent can asset accumula-
tion be explained in terms of some of the current concepts of consumer
economics?

Questions such as these are considered in this study. They are ex-
plored by making use of asset and other information collected from a sample
of young couples married in the summer of 1968 in the cities of Peoria
and Decatur, Illinois, and interviewed approximately every six months since
that time. The data used in this study relate to the first five years
of marriage.

Based on these data, this paper explores a nuTiber of questions relating
to the determinants of asset position. One such question is whether
a concept of permanent, or normal, income provides a more effective expla-
nation of asset accumulation than current income. This would seem a
logical h>q'othesis since most asset accumulation is made with long run
objectives in mind, so that some more permanent concept than observed income
should be more relevant.

-1-

A second question is the effect of debt on asset accumulation. Although
this question has not been explored much in the earlier research, it is
a frequent item of discussion, namely, whether couples that incur large
amounts of debt are more likely to improve their asset position (particularly
net assets) than couples with smaller amounts of debts.

A third question relates to the effect of initial asset holdings on
later asset accumulation. Is it true, for example, that the "rich get
richer" from the very start of the marriage?

A fourth question is the extent to which variables other than socio-
economic help explain variations in savings. To what extent, for example,
do attitudinal variables help in this regard? What about other variables
that may reflect asset accumulation tendencies? Two such types of variables
are available for this study, namely, attitudes toward savings (including
plans if any for saving), and measures of ownership of credit cards and of
durable goods.

In the case of socioeconomic characteristics, does it make any dif-
ference in explaining savings behavior whether one uses the characteristics
of the wife or the husband? This question rarely arises in other studies
partly because the famly unit is assumed to be homogenous in most economic
theory and partly because usually only one set of data is available any^'ay
(invariably for the husband). In this study, however, information on such
key characteristics as occupation and education are available for both
members of the couple. In addition, a set of questions was asked that led
to a classification of one or both members of the couple as the "family
financial officer." This raises interesting possibilities for exploratory
work, which will be discussed later.

-3-

These questions will be studied with regard to both gross and net
assets for two time periods — the first year and the fifth years after
marriage, the periods for which financial data are available.

As a basis for specification of alternative models, we begin by reviewing
some of the previous work in this subject area. The data set and its char-
acteristics are described in Section III. The presentation of the general
analytical approach, in Section IV, is followed by the results in Section V.
A concluding section reviews and discusses the implications of the
findings .

II. Earlier Studies

Until recent years, the principal attention given to consumer assets
in the economics literature was to their role as an independent variable,
partly as an input into the investment stream and partly as a determinant
of consumption and saving behavior. The general tendency in the classical
literature has been to treat consumer assets as an exogenous variable,
determined by savings accumulations over many past periods and, hence,
as a given quantity in explaining some aspect of consumer spending or saving
behavior in the current period. The principal exceptions seem to have
been Irving Fisher's treatise investigating, in part, the influence of
interest rates on asset accumulation,* and disucssions such as those by
Keynes on the reasons for holding particular types of assets.

It is only in the last 20 or 30 years that much attention has been given
to seeking to explain the level and, composition of consumer asset holdings.

*Fisher, I., The Nature of Capital and Income. New York: The Macmillan
Company, 1960.

i.w io i>moe

..9V(

-4-

partly perhaps because of the increasing amounts of such assets and partly
(and not unrelatedly) because of the increasing amount of data available
on the subject. Even so, empirical treatment of the determinants of total
asset holdings or of net worth are few and far between, and models of the
asset accumulation process in the early stages of the life cycle seem to be
nonexistent. Indeed, on a cross-section basis, which is of primary rele-
vance to the present study, only two such studies can be cited, both based
on the 1962-63 Federal Reserve Survey of Family Financial Holdings and focusing
on determinants of net worth. Thus, both Crockett and Friend,* and Projector
and Weiss,** found that income, age and various other socioeconomic variables
affected net worth holdings and that the long-run normal income elasticity
of net worth tends to exceed unity.

A number of other studies have focused on individual assets and still
others on determinants of assets divided into general categories. These
latter studies could be construed as attempting to exp. lin gross asset
holdings insofar as the individual categories sum to a meaningful total,
and it would therefore seem useful to summarize briefly the principal such
studies.

By running regressions with a different asset holding as dependent in
each case. Watts and Tobin, using the data from the 1950 BLS Consumer Expendi-
tures Study, concluded that "households tend to maintain some sort of balance
in their capital accounts both between assets yielding direct service and

*Crockett, J., and Friend, I.,. "Consumer Investment Behavior." In
Ferber, R. , ed. , Determinants of Investment Behavior. New York: National
Bureau of Economic Research, 1967, pp. 15-127.

**Projector, D.S., and Weiss, G.S., Survey of Financial Characteristics
of Consumers. Washington, D.C. : Federal Reserve Board, 1966.

-5-

financial assets, and between liquid funds and liabilities."* They, as
well as Guthrie at a later time,** found that as households moved \sp the
economic scale, more of all kinds of assets were acquired and debts were
reduced.

This absence of substitution among assets was also observed in a study
involving pension contributions by Cagan*** and in a study of a different
set of data by Henry Claycamp.^ Based on that study, Claycanrp propounded
a so-called "independence hypothesis," that "the aggregate distribution
of assets... approximates that which would be found if the ownership of
assets were independent. "##

Studying the demand separately by ordinary least squares for four
assets (marketable bonds, life insurance reserves, time deposits in com-
mercial banks and time deposits in other institutions). Hamburger found
interest rates and total wealth to be highly significant, income to have
negligible effect and some of the assets to be close substitutes for each
other. ### The latter finding was also obtained by Darby. ####

*Watts, H.W. and Tobin, J., "Consumer Expenditures and the Capital
Account." In Friend, I., and Jones, R. , Eds. Proceedings of the Conference
on Consumption and Saving, Vol. 2. Philadelphia: University of Pennsylvania,
1960, p. 48.

**Guthrie, H.Vi'., "Consumers' Propensities to Hold Liquid Assets," Journal
of the American Statistical Association, Vol. 55 (Sept. 1960), pp. 469-90.

■ ***Cagan, Philip, Pension Plans and Aggregate Saving. New York: National
Bureau of Economic Research, 19

#Claycamp, H.J., The Composition of Consumer Savings Portfolios.
Urbana, 111.: Uiiiversity of Illinois, Bureau of Economic and Business Research,
Studies in Con-iuraer Savings, No. 3/ 1963.

##Ibid, p. 54.

###Hamburger, M.J., "Household Demand for Financial Assets," Econometrica,
Vol. 36 (Jan. 1968), pp. 97-118.

####Darby, M.R., "The Allocation of Transitory Income Among Consunsers'
Assets," American Economic Review, Vol. 62 (Dec. 1972), pp. 928--41.

-6-

Broader models of asset-demand functions derived from utility theory
have been studied by Motley and Wachtel. Motley found, like Hamburger,
considerable interdependence among assets, and with both permanent income
and transitory income affecting asset holdings.* Wachtel used four cate-
gories of assets also, but two of these (durables, and consumption excluding
durables) were not financial assets. Using a partial adjustment approach,
he finds that transitory income more than permanent income influences these
asset holdings in addition to the lagged effects of the holdings themselves.**

Summing up this section, this past work would seem to suggest that a
large number of cross-section variables are likely to affect total asset
holdings, at least one of which is likely to be some concept of permanent
or normal income.

III. Descriptive Aspects

The data used in this study are from a panel of couples married in
the summer of 1968 in the cities of Peoria and Decatur, Illinois. The
husband had to be 30 years of age or less at that time and involved only
first marriages. These couples were interviewed approximatley every six
months since the fall of 1968, and a sizable amount of data were collected
relating to their money management and financial behavior. More specific
to the purposes of this study, a complete financial portfolio was obtained
in the third interview, when the couple had been married approximately one
year, and another complete portfolio four years later. The analytical
focus of the study is, therefore, on the characteristics and determinants

*Motley, B. , "Household Demand' for Assets: A Model of Short-run
Adjustments," Review of Economics and Statistics, Vol. 52 (Aug. 1970), pp.
236-41.

**Wachtel, P., "A Model (jf^f Interrelated Demand for Assets by Households,"
Annals of Economic and Social Measurement, Vol. 1 (April 1972), pp. 129-40.

-7-

of financial asset holdings as of the end of the first and fifth years of
marriage .

It is unfortunate that financial portfolios were not obtained as of
the time of marriage. From a survey point of view, however, it was felt
that such an attempt would be too damaging to the cooperativeness of the
panel members in view of the highly sensitive nature of this information.
The response rates obtained were quite good : approximately 72 percent of
the initial 313 couples were still in the panel after five years, which
provides a better base for analysis.

The other data used in this analysis refer to various characteristics
and attitudes of both members of the couple. In most instances, these data
were obtained for each member separately. The specific variables used are
best described in the later sections.* Here, we examine the characteristics
of the asset holdings and how they have changed over this period.

A general picture of the distrubtion of the couples by their total
financial holdings is provided in Table 1 for three main quantities, namely,
gross assets, total debts and net assets. Not surprisingly, the table shows
that in terms of gross assets most couples had relatively little at the
end of the first year of marriage (1969). More than half of the couples
had less than $5,000 in gross assets and an even larger proportion had
debts amounting to this much, ks a result, nearly half of the couples had
net assets that were either negative or negligible. Only 10% of the couples
had gross assets of $25,000 or more, but hardly any had net assets this
large.

*A more complete description of these data is available in a brochure
obtainable from the Survey Research Laboratory, University of Illinois,
Urbana, Illinois 61801.

-8-

1 . '^ex'cont Distribution of Families by Overall Finoncigl Holdings,
One Year av.d Five Vears After Marriage

Years arter marriage

Arrcun c

O-s 999

=1,000- 4,999

5,000- 14,999

15,000- 24,999

2 5,000 or more

Total

One

Gross Assets

22.5%
23.9
18.5
19.7
10.4
100.0%

Five

12. S%
13.7
36.1
16.9
20.8
100.1%

Total Debts

0-$ 999

31,000- 4,999

5,000- 14,999

15,000- 24,999

25,000 or more

Total

29.4%
28.8
22.0
19-8

100.0%

10.9%
7.3

24.4

36.8

20.7

100.1%

Net Assets

-S 10, 000 or less

-1,000 9,999

995- -999
1,000- 4,999
5,300- 14,909
15,000 or more
Total

18.6
29.7

30.2
16.9
4.7
100.1%

25.1%
22.4

13.7

15.8

14.8

8.2

100-0%

ncse in all cases is bcb.v'een 170 and 153 families.

-9-

Five years later, the situation had changed drastically. In terms
of gross assets, the distribution had shifted sharply to the right. The
proportion having gross assets under $5,000 had declined from over half
to just about one-quarter, while the proportion having gross assets of
$25,000 or more had doubled, to 21%. An even more pronounced shift to
the right took place in the debt position of the couples. Those having
debts unddr $5,000 declined from nearly 60% to 18%, while those having
debts of $15,000 or more tripled, from barely 20% to nearly 60%.

As a result, the effect on the net asset position was to flatten the
distribution markedly, with more couples having both less assets and more
assets. Thus, whereas no couples were in the red in terms of net assets
to the extent of $10,000 or more after one year of marriage, one-fourth
of the couples were in this rather precarious position after five years
of marriage. On the other hand, the proportion of couples having net assets
of $15,000 or more had increased from 5% to slightly over 8%, and couples
in the negligible net asset position had dwindled from 30% to 14%.

It is of some interest to note that, on the basis of a more disag-
gregative analysis, the only assets fairly common among these couples at
the start of the marriage were checking accounts, savings accounts in banks,
life insurance and a home; the frequency of the latter is not too surprising
in view of the relative scarcity of apartments in these two smaller-sized
cities. On the debt side, nearly 60% had a loan on a car; almost that many
owed something on other personal property; and nearly 40% had a home mortgage.
The frequency of ownership of these assets changed little over the five year
period, the principal characteristic being an increase in the frequency

-10-

of ownership o£ a home and of common and preferred stock. Particularly-
relevant for the later models is the fact that home owiers had much more
(gross) assets and debts than non-horaeowners, as might be expected.

Overall, the total assets of. these couples increased substantially
during these five years. However, the same was true of their debts, with
the result that their net asset position improved in some instances but
worsened in others. As a rule, the couples that were in the best financial
position at the start of marriage maintained that position, and correspondingly
for those that were less well off. This is brought out in Table 2, which
compares changes in total assets and in net assets between these two periods.
As is evident from this table, of those who had less than $1,000 of total
assets after the first year of marriage, 35% were in the same category,
after five years and another 25% had moved only into the next higher cate-
gory. In contrast, of those with over $25,000 worth of total assets after
the first year of marriage, 56% had this much assets after five years and
all of them had assets of at least $5,000 at that time.

A similar relationship is evident from the second part of Table 2
which relates to net assets. Thus, of those having significant negative amounts
of net assets (in the red by more than $1,000), nearly two- thirds were in
the same position after five years, whereas this was true of only 14% of
those having net assets of over $15,000 after the first year of marriage.
At the same time, of those having over $15,000 worth of net assets after
the first year, nearly 30% were in the same position five years later and
an equal percentage had between $5,D00 and $15,000 in net assets.

• 11-

ferceni- u-
for

Given Level of Assets

in Year 1
: 1

A. Gross Assets

Amount in Yeai

Amount in
Year 5

0-
$1,000

$1,001- $5,001-
5,000 15,000

$15,001-
25,000

$25,001
or more

0-$1,000

35.0%

6.8% 3.2%

3.8%

0,0%

1,001-5,000

25.0

13.6 9.7

5.9

0.0

5, 001-15, (

300

' 27.5

45.5 45.2

38.2

18.8

15,001-25:

,000

10.0

20.5 9.7

23.5

25.0

25,001 or

2.5

13.6 32.3

23.5

56.3

Total

100.0%

100.0% 100.1%

99.9%

100.1%

Ease

44 31

B. Net Assets

Amount in
Year 5 .

-$20,000 or less

-$1,001-
-10,000

3.6%

Amount in Year 1

$1,000-
-1,000

5.8%

$1,001- $5,001-
5,000 15,000

14.0%

0.0%

$15,001
or Etore

0.0%

-$10,001 to
$20,000

32-1

21.2

8.0

11.1

14-3

-$1,001 to
-10,000

28.6

28.3

16.0

14.8

0.0

$1,000 to
-1,000

17.9

15.4

18.0

11.1

14.3

$1,001 to
5,000

3.6

11.5

22.0

29.6

14.3

$5,001 to
15,000

14.3

15.4

12.0

11.1

28.6

$15,001
or more

0.0

1-9

10.0

22.2

28.6

Total

100.1%

100.0%

99.9%

100.1%

Base

-12-

IV. Analytical Approach

As noted in the preceding section, total assets can be represented
in either gross or net terms. The simple correlation betvveen gross
assets and net assets was in Year 1 and in Year S. Still, the two
terms are by no means equivalent, and there is no question that net assets
is a better measure of financial position than gross assets. Hence our focus is
primarily on testing the ability of various hypotheses to help explain
fluctuations in net assets. At the same time, there is also considerable
interest in the extent to which different hypotheses help explain variations
in gross assets and in debt, partly because these are of key interest in
themselves and partly because a more meaningful explanation of fluctuations
in net assets may well be obtained through first explaining fluctuations in
these other two variables. For this reason, we adopt a twin approach of
seeking to explain net assets on the one hand as the difference between
separate functions for gross assets and for debts and, on the other hand,
directly by expressing net assets as a function of alternative hypothesized
relevant variables.

By the indirect approach, we have a set of three equations, one an
identity expressing net assets (NA) as the difference between gross assets
(GA) and debts (DT) , and two behavioral equations, one for GA and one for
DT. By the direct approach, v/e have a single behavioral equation for NA.

For the explanatory variables, in addition to the three dependent
variables which may influence each other (not to mention lag effects), we
have the following four sets of variables:

■13-

1. A measure of family income which may be reported income

for the particular year (Y) , or a measure of long run, or "normal"

income fY ) .
^ n

2. A set of socioeconomic characteristics (SE) which includes the age,
occupation and education of the husband and wife separately, as
well as a variable identifying the "family financial officer."*

3. A set of variables reflecting the budget plans of the family and
the priority accorded to savings (AT) . The two key variables are
attitude toward savings and presence of a plan for purchasing goods
and making other expenditures .

4. A set of variables reflecting ownership of a home (H) and ownership
of various other financial instruments (OF) . These include purchase
of durable goods, number of major durables owned, and number of
credit cards owned.

Going back to the studies reviewed earlier and to the overall review

in the preceding sections, the following two general models are formulated

to explain fluctuations in net assets. For the indirect approach,
we have:

(1.1) NA « GA-DT

(1.2) GA = f [Y, DT, SE, AT, OF)

(1.3) DT = f (Y, H, SE, AT, OF)

*Such identification was made on the basis of answers obtained to three
questions relating to who paid the bills, who looked after excess funds
and who made decisions on major purchases. On the basis of these answers
it was found feasible to identify the family financial officer as the husband,
wife or both, separately in Year I and in Year 5.

-14-

In other words, xve test all four sets of explanatory variables in
both the GA and DT equations. In addition, since debts enter to a large
extent in the formation of gross assets, that variable is used as explana-
tory in Equation 1.2. At the same time, since the preceding section brings
out that most of the debt of these couples is related to the acquisition of
a home, ownership of a home is included as a dichotomous explanatory var-
iable in Ec^uation 1.3.

For the direct estimation of NA, two forms are used, namely:

(2) NA = f (Y, GA, SE, AT, OF)

(3) NA = f (Y, D, SE, AT, OF)

Once again, the four sets of explanatory variables are included in
each case. The difference between the two equations is the inclusion of
gross assets as an explanatory variable in Equation 2 and debts as an ex-
planatory variable in Equation 5. This is done to explore which of these
two indicators seems to affect net assets most strongly and, also, to
ascertain the extent to which net assets is influenced by debts.

The test of these two alternative approaches is based not only on the
goodness of fit and the significance of the coefficients but also on the
ability of each approach to estimate more closely the actual net assets of
the sample families.

Another dimension to the analysis is provided by the availability
of data for two periods — Year 1 and Year 5. As a result, estimates of these
models can be made for each of the periods separately and, in addition, a
further test can be made by seeing how these models estimate change between
these tvifo periods.

-15-

In each case we seek answers to the four questions raised in Section I
on determinants of asset position, namely, whether a concept of normal
income is more effective than reported income, the relevance of total debt,
the effect of initial assets and the role of variables other than socio-
economic. Also, for the socioeconomic characteristics, does it make any dif-
ference if they relate to the wife or the husband? In the latter case,
the test is made by using three alternative formulations of the socioeconomic
characteristics, namely, only those of the wife, only those of the husband,
and neither.

All parameters were estimated by ordinary least squares using linear
forms, with the dollar variables (NA, GA, DT and IN) in arithmetic terms.*

V. Results

Parameter estimates obtained by applying to Year 1 data the foregoing
models incorporating the alternative variations of the income and socio-
economic variables discussed earlier are presented in Table 5. For the socio-
economic set, three variations were. tested, one containing variables reflecting
only the characteristics of the husband, one with only the characteristics
of the wife, and one with the characteristics of neither. The "normal"
income of the family was estimated as a linear function of the age, education
and occupation both of the wife and of the husband, of home ownership, and of
occupation both of the wife and of the husband, of home ownership, and of
savings attitudes of the couple; these were the variables felt most likely
to reflect the longer run level of family income. Considering that this
was the first year of marriage, and that many if not most of these couples
had not yet had a chance to establish a clear career path, the validity

*Since there were an appreciable number of zero values, especially for
Year 1, and since net assets were frequently negative, expressing these
variables in logs was not feasible.

■16-

^0|

I r-( i I
I'll
O

•T3

I I I I I
I I t t )

<
I "^

• I I i I

*
*

¥
O

VTi

o
I

{SI

1 *^

r-l 1 1

1 •

• 1 1

*
o

*
*
CM

"^i-l

1 .—1
1
O

\ 1
) 1

to|

1 i-H

!-( 1

1 1

</>

1 ■

• 1

1 1

•1-1

(ft

cm!

1 1

r-t 1

1 1 !

1 1

• 1

1 1 t

i-il

1 1

i-i 1
• 1
o

0.1

e
o

«J

c
o

o
u

tA
0>

OS
Q

<1>

tn
o

1-*

•

•a

•

■«•

1 N^

r-t

t •

•H

Irt

) 1

1 1 I~0

1 1

o
tn

(]>

1 CM

»-t

. 1

•

■X) (U

c >

•

/ — »

CO 0)

.— 1

v-_^

lO o

■H

^-v ■ -O

/ — V

O 'H

^— -,

'-^^ (D

•>

« »

C "tj

3 >— '^-^ .-<

•^/-^

o o

rt 3

+->

\^-/

^-^ X /— ^ r-H

^ +J

• ^— '^ .H

* 4J

-a

CL.-H

C f4 >-/^

J^ /— >

f-t C!j

■U

O tJO ^ 10

bo^ 3

<U 4-i

/ — \ 1 — ^

' — \< — > -H

•H S rt -d C

S rt ^

. a)

<— )

to 03

x: s

r-<

x: 3 -p

■(-> 1 O 0) 3

1 o

o u

•*— / v—/

WW c(J

«J • -H r-f 1

. .H f<

<u •

O tw f^ ^ .H

U-i M O

iU CTj

•H

Pu (U

o o

c; <D 3

3 O O -H e

o <a xi

X tJ

X >

IL U.

-r^

bO ClCO

--C M --I ^ (D

$-4 ^ rt

oF <4-l

UJ <

a. ci-

< < M

tU CU O C/D tn

a. u »J

E-

• •

•

• • •

• # •

•

•f*%

•H

r-i

CM M

>* tn

■•£)

00 CTi O

.-( CNI t^ ■^ LT)

\0 i^ 00

CTi

t-H

•H

r-<

rH f-H I—I .-H .—1

.-H 1— t 1— <

r-f

CsJ

tft

-17-

of these variables for this purpose is unclear, but are used any^vay for
later comparison. At the same time, since the socioeconomic char-
acteristics are major components of the normal income estimate they
are not used as separate variables in the functions including normal
income .

To bring out more clearly the principal results, parameter estimates
are given in Table 5 (as well as in later tables) only where they are
significant at the .10 level or beyond. A variable that was included in
a function and is not significant at that level is shown by a dash. A
blank in a particular space means that the variable was not included in that
function.

Looking at the results in Table 5, we find that the adequacy of the
different models varies substantially both with the model and with the
dependent variable- All of the gross assets functions explain approximately
10 percent of the variation in that variable, the total debt functions
seem completely ineffective, while the net assets functions explain about
30 or 70 percent of the total variation, the latter when gross assets is
included as an independent variable. With respect to the four questions
raised initially on the importance of different factors influencing asset
holdings, the normal income concept used seems to have about the same rele-
vance as reported income in explaining asset holdings. It is highly signifi-
cant in the gross assets function, is also significant in one of the net
asset functions though it is not statistically significant at the , 10 level
in the total debt function. On the other hand, reported income is also
statistically significant at the .01 level in the gross assetsand net assets
functions.

-18-

Total debt is not important in explaining gross assets, but does have
a strongly negative significant influence on net assets. Although gross
assets seems to be a more important influence on net assets than total
debt, these results provide a clear indication that at least at the start
of the marriage couples with large debt tend to have small net assets, and
conversely. This is undoubtedly due to the need of many couples at the be-
ginning of a marriage to borrow money to furnish living quarters and,
occasionally, also to buy a house.

Other variables appear to be of lesser importance. Except for income,
the socioeconomic variables seem to be of little importance. However,
gross and net asset holdings seem to be affected positively by attitudes
toward saving and negatively by the husband being the family financial
officer.

It might be expected that fitting the same functions to data for Year 5
would yield better results than were obtained for Year 1, because after
five years of marriage the couples would have had the time to better estab-
lish their life styles, so that patterns of asset accumulation and determinants
of this accumulation would be more apparent. This is indeed the case, as
is evident from Table 6, which presents data for Year 5 corresponding to
those given for Year 1 in Table 5.

Perhaps the most obvious result is that except for some of the net
assets functions, all the coefficients of determination for Year 5 are con-
siderably higher than the corresponding figures for Year 1, and this time the
debt functions are statistically significant, at the .01 level. For Year 5

■19-

•P
O

rj|

tr.
tn
O

to]

*
00

O I r

• I I
o
I

I • I
o

r I
I I

fol

1 1

0.1

t/1

</)

tNI

O I

1 1 t

•M

- 1

1 ) 1

<1
1^

-h|

I 1
1 1

1 1 fH
1 1

* *

at t-

O 00

« ■

•

o o

I--

f-t

1—1

4->

•

in
«>

a *

00 00

O CO

r>-

• •

o o

•15,

•?-

«M

1 r-1

•

irt

•

(-1

* < <

r- •^ n-

f-H t ! I i~l I— I

•ill • •

o o o

1 1 1 1 o

1 1 1 1

I I
I I

••

r-l

•

■it

t-H

«>

*<1

fi^

LTi

1-1

tfi

•

•■-(

■it

•*

■St

•*

CSJ

t-l

-H

I— (

fvj

r-(

•

o
1

•»t

«-«

■x

•y

?■

Lft

t) <D

/— \

§5

.£

.-(

<*-!

■~r'

in o

/— ^ -C3

/ — \

O —1

/— V

/ — >

x: «■>

• •

c -o

~ ' /— » r-l

V ' f — s

o o

rt 3

^ '-^ --I

T-4 ■(-;

1/1

■ ^-^ ^ •'-1

. <~J

-. 4->

•n

PLi-H

^^ >-'-^

U /-v

.-( rt

■t->

t/1

bo—) V)

&0r-< IS

O 4->

^~^

, — ^

^-, ,-^ -H

■H

E c« TJ C

5 oj ^-^

• (U

m ri

•-H

x: s 4->

■p

1 O O 3

1 o

o o

0) ■

% — '

Cu O

t)-4

<u <a 3

o a> -H E

O <D J=

.-)

r— (

»— *

.c o

X >

U-

■H

&0 bed

'H .-1 ^ (U

f-i -H rt

<£

LO LfJ

UJ <

tu.

U-

< < cu

a. u to w

Cm U J

>■

«:

* •

• • • k

. .

•

r-i

{M tri

t--

CO CT) O

.— t

CM to 'd- Ln

so t^ 00

1—1

(NJ

VJ-

L/*j

■o

-H

>— 1

.-1 t-H cH r-H

i-H r— 1 r-i

0-1

r-i

O-J

- <

-20-

nearly 60 percent of the variation in gross assets holdings is explained
by the various sets of independent variables. Total debt is now a highly
significant explanator>' variable. Both normal and reported income are
separately also highly signi£icant,*but neither is as important as total
debt .

Either reported income or normal income is highly significant
in accounting for variations in total debt. Also important in explaining
total debt holdings are a detailed expense plan, and the
stock of durables, both with expected positive signs, and wife not working.
In other words, total debt tends to be higher if the family has a detailed
expense plan, if the family has a large stock of durables and if the wife
is not working.

The most important influence on net assets is seen to be the level of
gross assets, as was the case in Table 5. Net assets in Year 1 is highly
significant when the gross assets variable is not included, but is still
statistically significant (but much less important) even with gross assets
in the equation. Total debt is highly significant, as is reported income,
though gross assets is clearly the dominant explanatory variable. Especially
interesting is the fact that in Year 5 total debt and net assets are now posi-
tively correlated, suggesting that those who borrowed much money initially
seem to have made good use of the funds.

Turning to the four questions asked initially about influencing var-
iables, this time reported income seems more important than normal income.
This is evident for all three t>T)es of assets. As in Year 1, total debt

*In addition to the variables included in the normal income specification
mentioned earlier for Year 1, this estimation function for normal income in
Year 5 included reported income in Year 1. Even so, the goodness of fit
was not much better than for the Year 1 function.

-21-

is a major influence, this time for net assets as well as for gross assets
with, as noted previously, the relationship with net assets being the oppo-
site to that observed for Year 1. Bearing out the earlier observation
that in the first few years of marriage the couples would be adjusting their
life styles is the fact that the initial holdings of these assets or debts,
while showing some positive relationship to the Year 5 holdings, do not
seem anywhere near as important as other variables.

As in Year 1, the socioeconomic variables show only scattered influence.
The principal such influence is the tendency for total debt to be less if
the wife is working.

Unlike the results in Table 5, savings attitudes seems of little impor-
tance in explaining variations in these holdings, but stock of durables
retains some importance, particularly for debt and net assets.

The results using the same models to explain changes in assets and
debts from Year 1 to Year 5, in Table 7, show that these functions explain
about half of the variation in the change in gross assets, about a fifth
of the variation of the change in total debts, and either about 10% to 70%
of the variation in the change in net assets, depending on the exclusion
or inclusion of gross-assets change as an independent variable.

Essentially these results are not too different, in terms of significant
variables, from those for Year 5 in Table 6. Thus, the principal explana-
tory variables for the change in gross assets is the change in total debt
and the level of reported income. Reported income also enters into the
explanation of change in total debt and of change in net assets. In the

-22-

•*!

fvl

1—*

! 1

1 •-<

•

1 1

<3 <1

1 "-I

r-l

r-H 1

• 1

■K

1 1

1 r-<

1 1

1 •

o
1

< *

•* 00

>Ji

■«.f

cmI

1--* rW I

1 r— 1

i-f

• !

■

•

o o

E-

<J *

*<3

eo CO

O <*

r-ll

r~{ r-l i

t T

■— (

K) rH

• • 1

1 1

•

<
<

'^l

CJ|

o o

o o o

<3 <]

r>) r^

1 I 1-1 i~«

.-^ 1

1 t • •

• 1

o o

1 1 1 i i-H I 1

(ill • 1 1

o
s

fvj

111 .-Hill

1 r^

III •III

1-1

Irt

*
*

o
o

-tc

f-*

Ifi

r-.

•

iJ)

«-<

fvj

t-H

«-(

1
1

«
o

•H
U

♦

•

••a

■*

•o

tS-

-a

{

1— 1

•

y-i

•H
C

->:

• r4

r-^

•Ji

r-i

•

«5H

■r-i

r-H

irt

t-I

o •

M
■H

t3 O

«A

C >

•

•X3

/— \

* (U

—(

•H

tri O

cii

.-— ^

-o

/ — \.

O '-I

x: '

c -a

>-<

•r-i

v — ' -■ — ».

.—1

^-^ /— *^

o o

OJ 3

-CJ

,- — ^

r~i

e-i ^

■r-i

. v.—'

-> JJ

-d

CL-H

•H

J-4

'^ — '

f^i ,-^

^ «

+->

P-i

/—>

,---^

M r-!

CiOr-i 3:

O 4->

**— \

^ — ^

/ ^

y— ^

e rt

6 ctJ ^

• «

in rt

.—4

<4-(

V— /

! O

1 o

o y

s_>

>_ /

• -H

• -H ^^

<D •

CvS

•

■^ u

i-i

•l-l

M-i 5^ O

® CO

Pw <D

<+-!

o a>

•H

O O Xi

x; u

X >

•H

•o

J^ rH

!-i .-i c3

'■JD

\-n

l>J

H -H

tu <

LL,

• m

Ci. u

a- u -J

• •

B *

!• » e

•(->

r-<

CM K)

■<*

(—4

t^ -^^f

l^ 00 CTi

.-1

•H

e-4

T-H

(-^ r— 1

r—f

«— < .-H <-(

rsj

<Ni

-23-

latter case, change in gross assets is again by far the most important
variable.

In terms of the questions about the four types of separate influences,
once again it appears that reported income rather than the normal income
concept used is the more relevant. In a sense, this is not surprising since
even after five years, it is not clear whether normal income estimated on
the basis o£ regressions between actual income and a set of socioeconomic
characteristics is likely to be very "normal" from the point of view of
reflecting saving behavior. There is no question that many of these couples
were still getting adjusted (and a few getting divorced) , so that not many
of them are likely to have reached a level of equilibrium after five years
to lend much validity to a measure of this type.

In terms of change, total debt now has no relationship to net assets,
though it is, as might be expected, highly correlated with gross assets.
Socioeconomic variables once more do not exhibit much importance except
for the wife working, which tends to reduce the am.ount of debt. Other var-
iables do appear to have some effect, especially ownership of credit cards
(which tends to increase debt and to decrease the change in net assets),
presence of a detailed expense plan (which also tends to raise the change
in debt and loiver the change in net assets), ownership of a house (which
increases debt), and change in stock of durables (ivhich tends to bring
about changes in a similar direction for all three types of assets).

Is the direct or indirect approach better for explaining net assets?
The answer from these data, using all three models, is provided in
Table 8. This table provides estimates of the goodness of fit obtained

-24-

Period
Year 1
Year 5
Year 1-5

8. Goodness of Fit of Alternative Approaches to
Estimating Net Assets

Dependent

NA estimated
as GA-fb

NA directly

variable

(2) GA
0.73**

(3) TD

0.13**

0.36**

0.21**

0.80**

0.27**

0.08**

0.76**

0.16**

-25-

by the two different direct models (alternately excluding and including
total debt in the net assets function) and for net assets estimated as a
difference between estimated gross assets and total debt.

As is evident from this table, the direct approach seems to yield much
better results, especially when gross assets is included in the net assets
function. These results serve to confirm those evident from the preceding
three tables, which brought out the importance of gross assets as an explana-
tory variable in the net assets function, and which demonstrated that this
function had a much higher goodness of fit than any of the gross assets
or total debt functions. In this case, in other words, disaggregation of
net assets into its principal components is not likely to yield any improvement
in explaining fluctuations in that variable.

VI. Conclusions

To come back to the questions raised at the beginning of the study,
some fairly definitive answers are indicated by the foregoing results, at
least as applied to this restricted data set. For one thing, the tables
in Section III as well as the regression results suggest clearly that, at
least in the first few years of the marriage, those who start out with more
tend to maintain and, if anything, widen the margin. Not only are the
autocorrelations between Year 1 and Year 5 for gross assets and net assets
strongly positive but the dispersion of these asset distributions increases
markedly over time.*

*For example, the inter-quartile range for net assets increased between
Year 1 and Year 5 from $2,100 to $11,100; the corresponding figures for
gross assets are $2,400 and $20,300.

-26-

Which variables differentiate between couples that improve their
financial position in these first few years of marriage and those that do
not? A strong base established by the time of marriage is an obvious answer,
as suggested by the previous point. In addition^ the results offer the
intriguing suggestion that the couples best off financially after five years,
both in terms of gross assets and of net assets, are those who are venture-
some enough to acquire substantial amounts of debt. Thus, the primary
determinant of the amount of gross assets after five years, as well as of
the change in gross assets during the first years of marriage, is the amount
and the change in total debt, respectively. In turn, the most important
variable explaining the level of net assets after five years, as well as the
change in net assets, is gross assets. For both types of functions, the
influence of debt on gross assets, and of gross assets on net assets, is
much greater than the autocorrelation of gross assets or net assets with
themselves.

Unelss it be inferred from these findings that the way for a young couple
to get rich is to rush into debt, it should be stressed that the positive
influence of debt might have been the result more of general economic condi-
tions than of the financial accumen of these couples. This is be-
cause the principal form of debt was represented in this sample by purchase
of a home. About 43% owned a home already at the time of marriage, and this
percentage had increased to 82% after the first five years, a tendency not
unusual in smaller cities like Peoria and Decatur. Since the period during
which they were making these purchases, 1968-73, was characterized by con-
tinually rising prices for homes, the debt that these couples incurred
to acquire homes was accompanied by continually rising equity that served

j3j-j

,oy »' in-x

.'jiiaor;

-27-

to raise the value of both their gross assets and net assets. Whether
the same favorable conditions would hold at other times is much more prob-
lematical.

These results also suggest that a concept of normal income is not
as useful for explaining differences in these holdings among different couples
as is current reported income. This is not too surprising in view of the
difficulty of imparting much meaning to "normal" or "permanent"
income at this early stage of family formation. That these concepts may
be more useful at a later stage is suggested by the fact that normal income
seems to be more likely to be significant for the asset functions in the
fifth year of marriage than in the first year.

With regard to socioeconomic variables other than income, it is rather
surprising to note how infrequently such variables appear to show any sig-
nificance. There is some indication that the presence of a working wife
tends to be associated with a smaller volume of debt and that the husband
being in a professional or managerial occupation is associated with more
gross assets, while the reverse (oddly enough) is true if the wife is in
a professional or managerial occupation. None of these variables are,
however, strongly significant.

More noticeable is the influence of various other financial as well
as some attitudinal variables on these asset holdings. Thus, a larger
stock of durables is associated with more gross assets and more net assets
after five years, even though the value of this stock does not enter into

-28-

these dependent variables.* A positive attitude toward savings also seems
to contribute to more assets, especially at the start of the marriage,
though this variable does not show up in the Year 5 functions, possibly
because its effect has by then been absorbed by the financial variables.
On the other hand, the presence of a detailed expense plan does not show
up in the Year 1 function but shows up clearly in the Year 5 functions, acting
to increase total debt and to depress net assets. Conceivably, such ex-
pense plans are developed only over time so that this type of question may
not be too meaningful when asked at the very beginning of a marriage.

There is further some tendency for assets to be less if the family
financial officer is either the husband or the wife rather than both jointly.
The husband as the financial officer seems to have some influence toward
decreasing the amount of assets, particularly so in Year 1.

In closing, it cannot be overemphasized that these results are based
on a limited data set and should be treated only as suggestive for future
work. Nevertheless, in view of the virtual absence of any studies of the
asset accumulation practices of married couples in this very early stage
of family formation, these results should provide a basis for more intensive
study in the future of this key segment of the population.

*In a sense, however, these durable stocks do enter indirectly, to
the extent that purchasers of a home are likely to also buy more durables
to equip that home.

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Asset accumulation in early married life"

See other formats