Full text of Working Papers (Federal Reserve Bank of Chicago) : First-Time Home Buyers and Residential Investment Volatility, Working Paper 2007-15

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Federal Reserve Bank of Chicago

First-Time Home Buyers and
Residential Investment Volatility
Jonas D.M. Fisher and Martin Gervais

WP 2007-15

First-Time Home Buyers and
Residential Investment Volatility∗
Jonas D.M. Fisher
Federal Reserve Bank of Chicago
jfisher@frbchi.org

Martin Gervais
University of Southampton
gervais@soton.ac.uk

November 29, 2007

Abstract
Like other macroeconomic variables, residential investment has become much
less volatile since the mid-1980s (recent experience notwithstanding.) This paper explores the role of structural change in this decline. Since the the early
1980s there have been many changes in the underlying structure of the economy,
including those in the mortgage market which have made it easier to acquire a
home. We examine how these changes affect residential investment volatility in
a life-cycle model consistent with micro evidence on housing choices. We find
that a decline in the rate of household formation, increased delay in marriage,
and an increase in the cross-sectional variance of earnings drive the decline in
volatility. Our findings provide support for the view that the “Great Moderation” in aggregate fluctuations is not just due to smaller aggregate shocks, but
is driven at least in part by structural change.
JEL Classification Numbers: E22, E32, J11, R21
Keywords: Business Cycles; Housing; Residential Investment; First-Time
Home-Buyers; Great Moderation
∗

We are grateful to Jay Zagorsky for giving us his NLSY net asset data. Thanks to Marco
Bassetto, Jeff Campbell, Cristina DeNardi, Henry Siu, participants of the 2006 LAEF Housing
Workshop and the 2005 NBER summer institute for their comments on earlier drafts of this paper.
The second author gratefully acknowledge financial support from the Social Sciences and Humanities
Research Council of Canada. We are grateful to Faisal Ahmed, Scott Brave, Nishat Hasan, Eric
Vogt, and Saad Quayyam for research assistance. The views expressed herein are those of the authors
and do not necessarily reflect those of the Federal Reserve Bank of Chicago or the Federal Reserve
System.

1

Introduction

Over the last twenty years macroeconomic volatility has declined significantly. Recent
experience notwithstanding, residential investment’s volatility has declined by more
than most other aggregate variables. Strikingly, the variance of de-trended residential investment from 1984 is a fifth of what is was before then. What underlies this
dramatic drop in volatility? We argue in this paper that lower rates of household formation, delayed marriage, and an increase in the cross-sectional variance of earnings
account for most of the residential investment volatility decline.
There are two possible sources of a decline in economic volatility: structural change
and changes to the variability of shocks affecting the economy. Obviously smaller
aggregate shocks without a change in structure might account for the decline, and
several authors have pointed to this possibility.1 Still, there have been many changes
in the underlying structure of the economy. In the housing market there has been
substantial regulatory change, the maturation of the secondary market for mortgages,
and a proliferation of new mortgage products. All of these developments have made
it easier to obtain a mortgage. Other structural changes with the potential to affect
housing volatility include a decline in the rate of household formation due to the end
of the baby boom, delayed marriage, and an increase in the cross-sectional variance
of earnings. This paper disentangles the contribution of these structural changes to
the decline in residential investment’s volatility using a quantitative life-cycle model
and evidence on housing choices from micro data.2
Our underlying hypothesis is that structural change has driven the dynamics of
residential investment through its impact on the behavior of first-time home buyers.
First-time home buyers, like all home buyers, are subject to various credit constraints
when trying to buy a house. Since first-time buyers are typically young and so have
relatively low income and savings, these constraints are likely to bind more often than
for older, repeat buyers. How residential investment responds to a given aggregate
1

See for example Stock and Watson (2002) and Ahmen et al. (2001).
We are not the first to examine the role of changes in financial markets on the business cycle.
Closest to our work is probably Campbell and Hercowitz (2004) who study the impact of financial
innovation on volatility in a two-agent infinite horizon framework with a durable consumption good.
Dynan et al. (2006) survey some of the literature which seeks to understand the impact of financial
innovation on aggregate volatility.
2

2

shock depends on the severity of the credit constraints and the number of home buyers
hoping to enter the housing market when the shock occurs. The structural changes
we consider affect both the severity of credit constraints and the number of potential
home buyers.
We analyze the behavior of home buyers in the National Longitudinal Survey of
Youth (NLSY) and the Panel Study on Income Dynamics (PSID). The evidence we
collect guides the formulation, calibration and evaluation of our model. We emphasize
three main findings. First, the decision to purchase a home for the first time is closely
associated with marriage. Second, in the years immediately preceding an individual’s
first home purchase, income and net wealth grow rapidly. Afterward, income is stable
buy net wealth continues to grow, but at a slower rate. Third, home ownership and
marriage occur later in the life-cycle for cohorts born later in the sample. The close
connection between marriage and home ownership suggests that the initial house
purchase is partly a response to a “family shock,” rather than being driven entirely
by wealth accumulation and portfolio allocation considerations. We incorporate this
feature into our theoretical analysis.
We disentangle the contributions of the various structural changes to the dynamics
of residential investment with a life-cycle model. The decision to buy a house is
influenced by an exogenous shock which drives up the housing subsistence level for
an individual. We assume that the larger houses needed to satisfy the induced increase
in housing demand must be owned. Individuals may borrow to acquire a house, but
they must satisfy a down-payment constraint to do so. We calibrate the model’s
stationary equilibrium to aggregate and microeconomic evidence. The calibrated
model reproduces the patterns of income and asset accumulation around the time of
the first purchase we find in the micro data.
To assess the impact of structural change on volatility, we simulate the dynamic
response of the model economy to exogenous shocks under the regulatory, demographic, and idiosyncratic earnings regimes of the pre- and post-1984 periods. Our
main finding is that most of the decline in volatility would have occurred in the
absence of any change in mortgage markets, due to a reduction in the bunching of
individuals near the threshold determining home-ownership brought on by a lower
rate of household formation, delayed marriage, and an increase in the cross-sectional

3

Table 1: The Decline in Aggregate Volatility After 1984
Variable

Variance Ratio

GDP
Residential Investment
Non-residential Investment
Consumer Durables
Non-durable Consumption
Services Consumption

.22
.21
.56
.18
.44
.63

Note: Variance ratio is ratio of variances computed for the sample 1984-2006 to the
sample 1948-1983. The quarterly data is first logged and then detrended with a band
pass filter excluding periods less than 6 months and greater than 8 years.
variance of earnings.
The next section describes some of the empirical evidence which motivates our
study. We then discuss our findings from the micro data. Section 4 describes our
model and characterizes its solution. Calibration and simulation follows and then we
conclude.

2

Motivating Evidence

This section discusses some of the evidence which motivates our study. We begin by
documenting the decline in volatility of various components of aggregate expenditures.
After this we discuss some time series evidence which suggests studying first-time
home buyers may be fruitful for understanding changes in residential investment
volatility. Finally, we provide a brief overview of structural change in the housing
market.

2.1

The Decline in Volatility

Table 1 illustrates the decline in aggregate business cycle volatility. The entries are
ratios of variances calculated using quarterly data over the samples 1984:I–2006:IV
4

and 1947:I–1983:IV. The table indicates what many others have noticed, that there
has been a broad-based decline in macroeconomic volatility. For example, GDP
volatility after 1984 is just 22% of its volatility before 1984. Of the other variables,
the volatility of durable consumption expenditures and residential investment have
fallen the most, slightly more than GDP as a whole and more than twice as much as
any of the other expenditure components.3

2.2

Why Study First-Time Home Buyers?

Of all the participants in the housing market, why focus on first-time home buyers?
That we focus on an aspect of the demand for housing is easily justified on empirical
grounds. For instance, while the consumption price of durable equipment is strongly
counter-cyclical, the consumption price of residential investment goods are strongly
pro-cyclical. Much has been made of the former as justifying a focus on productivity
disturbances in the investment goods producing sector. Similarly, the latter suggests
a demand side interpretation for the change in residential investment’s dynamics.
Fluctuations in residential investment are directly related to housing transitions
between and within tenure status. Based on our analysis of the PSID, transitions
between renting and owning and between owning one home and owning a different
home account for the largest numbers of households and are the most closely associated with fluctuations in residential investment. Figure 1 displays annual estimates
of these transitions from 1969 to 1997 based on calculations of ours using the PSID,
which are described in more detail in the Data Appendix. The units for the two series
are percentage of households.
The rate of transition from renting to owning is our gauge of the influence of
first-time home buyers on aggregate housing dynamics. We do not focus on first-time
buyers directly because of a measurement problem. Specifically, we need to observe an
individual over a long time period to detect a first-time home buying event. Observing
the move from renting to owning or owning to owning only requires two consecutive
years of data and so are easier to measure. To the extent that we are able to measure
3

Since the structural changes we focus on affect the demand for consumer durables because it is
in part derived from the demand for housing, our analysis should be informative about the decline
in durables volatility even though we do not model this explicitly.

5

0

Percent of Households
.5
1
1.5

2

Figure 1: Housing Market Transitions

1970

1974

1978

1982
year

Rent−to−own

1986

1990

1994

Own−to−own

Source: Our calculations using the PSID.

first-time buying, the majority of rent-to-own cases and its main fluctuations seem to
be due to first-time buyers.
We confine our attention to those transitions involving a purposeful increase in
the quantity of housing demanded. Either of the two transitions, rent-to-own or ownto-own, can result from an increase in the size of house demanded, say because of
marriage, or because of a move to a higher paying job. We assume that the demand
for housing rises when one of these two events occurs and in addition there is an
actual increase in the amount of housing consumed. Therefore we further restrict the
sample by focusing on those transitions which involve an increase in the number of
rooms in a household’s home.4
4

In the data, we measure those transitions associated with answering a survey question on why the
household moves by saying for ”productive” reasons or for ”consumption” reasons (where marriage
is stated explicitly as an example event.) The other categories are ”Involuntary” and ”Ambiguous.”
By excluding ”Involuntary” we leave out increases in housing demand due to household dissolution,
an event which influences the demand for housing which is not present in our model.

6

Two features of the time series plotted in Figure 1 suggest that first-time home
buyers may be important for understanding the change in residential investment
dynamics. First, the buyers making the transition from renting to owning make up
the largest fraction of households making a housing transition at any point in time,
and are always greater than repeat buyers who transit from owning one home to
owning again. Second, the variance of rent-to-own declines about 75 percent after
1984 compared to before, while there is only a 12 percent decline in the variance
of own-to-own. The hypothesis that there has been no change in variance for the
rent-to-own variable after 1984 is rejected at the 6 percent level of significance. For
the own-to-own variable, this hypothesis is only rejected at the 29 percent level of
significance. These differences are driven primarily by the two boom and bust episodes
in the earlier sample period which are much greater in amplitude for the rent-to-own
variable.

2.3

Regulation Q

Regulation Q has long been viewed as a source of volatility in housing, especially
within the Federal Reserve System.5 The regulation was a product of the Banking
Acts of 1933 and 1935. Initially, it prohibited interest on demand deposits and authorized the Federal Reserve to set interest rate ceilings on time and savings deposits
at most commercial banks. Until the mid-1960s the regulation had little effect on the
supply of mortgage funds because the ceilings were never binding.
The regulation began to exert substantial influence on the mortgage market after
1966. In 1966 the interest rate ceilings were extended to thrift institutions making
the regulation applicable to almost all mortgage originators. As interest rates rose,
the ceilings were binding for extended periods, as indicated by the shaded regions
in Figure 2. When the interest ceilings were binding, funds flowed out of deposits
to instruments with higher yields. In an era when mortgage lenders had limited
alternative sources of funds, the consequence of this outflow was a fall in the aggregate
supply of mortgage funds. Some mortgage lenders were also subject to State usury
laws, so instead of raising interest rates they rationed mortgages, often by lowering
5

See, for example, Gilbert (1995), McCarthy and Peach (2002), Jaffee et al. (1979), and Ryding
(1990). This section builds on Gilbert (1986) and Ryding (1990).

7

Figure 2: Regulation Q and Residential Investment
-5.8

-6.0

-6.2

-6.4

-6.6

-6.8

-7.0
1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Note: Solid line – residential investment per household. Shading – periods in which
interest rate ceilings were binding.
loan-to-value ratios. These considerations strongly suggest that Regulation Q had a
destabilizing effect. Indeed, as Figure 2 shows, the periods when the interest rate
ceilings were binding coincide with some of the largest fluctuations in residential
investment.
As lawmakers became convinced of its destabilizing effects, Regulation Q began to
be dismantled in the late 1970s. In mid-1978 depository institutions were permitted
to issue large denomination “money market certificates.” The Monetary Control Act
of 1980 was another key development. This Act set out a timeline for phasing out
interest rate ceilings and eliminated State usury restrictions on mortgages. Further
deregulation followed, with banks and thrifts permitted to offer deposit products to
compete with money market funds in 1982. By 1984 the interest rate ceilings were no
longer binding on thrift passbook accounts and Regulation Q ceased to be a major
factor in the mortgage market.6
6

During the 1950s and early 1960s there were other forces influencing mortgage supply which had

8

2.4

Innovations in Mortgage Supply

Since the early 1980s, as the the secondary market for mortgages has expanded dramatically and a much wider range of mortgage contracts have become available, it has
become much easier to purchase a home. Moreover, through specialization improved
efficiencies have lowered the cost of origination and servicing mortgages. It was understood early on that the mortgage market was undergoing fundamental change.
Indeed, in his introduction to a volume describing the changes, Florida (1986) wrote
“the past few years have witnessed a virtual revolution in the way this nation finances housing.” Deregulation, and innovations in information technology, credit
risk modelling and asset backed securitization are some of the factors which drove the
changes.7
There can be little doubt that these developments have made it easier for households to obtain mortgages. One way they have done so is by reducing the cost of a
given mortgage. For example, initial fees and charges for mortgages are now only a
fifth of what they were at their peak in the early 1980s.8 In addition, as described by
Van-Order (2000), the secondary mortgage market has increased the supply of capital
to mortgage markets. Lower transactions costs and a greater supply of capital have
driven down mortgage interest rates relative to what they would be otherwise.
The kinds of mortgage products now available have also made it easier to obtain a
mortgage. The typical mortgage before the 1980s involved the potential home buyer
effects similar in Regulation Q. At this time a large proportion of mortgages were guaranteed by
the Federal Housing Administration. These mortgages were subject to interest rate ceilings so that
non-price rationing would occur when market interest rates were higher than the ceilings. Another
factor was the elimination of World War II era regulations which kept loan to value ratios low. In
the aftermath of the Korean War these regulations were relaxed and a building boom ensued.
7
Florida (1986) contains several essays describing mortgage market deregulation. Gerardi et al.
(2007) provide a recent overview of how the mortgage market market has evolved. Edelberg (2003)
discusses the expanded use of sophisticated credit scoring methods in the mid-1990s. The secondary
market in Federally backed mortgages has been around since the 1930s. In 1970 the Federal government began sponsoring a secondary market in conventional, i.e. not Federally backed, mortgages,
but it was not until the 1980s that large numbers of mortgages were securitized. By 1991 the secondary market securitized more than 40 percent of outstanding home mortgages. Ryding (1990)
and Van-Order (2000) describe the evolution of the secondary market in more detail.
8
This is based on data from the Federal Housing Finance Board. Some of the decrease in measured
transaction costs may be due to the fact that in the late 1970s and early 1980s mortgage originators
used initial fees and charges to raise the price of a mortgage when usuary laws restricted their ability
to raise interest rates.

9

satisfying a relatively rigid set of criteria on the loan-to-value ratio, the mortgage
cost relative to income and other measures of creditworthiness.9 This rigidity can
be traced to the limited development of the secondary market for conventional mortgages and regulatory restrictions which put limits on the geographic area mortgage
originators could lend in. Since mortgage originators had few options to hedge against
their risk, they were forced to restrict the amount of risk they could take on with any
given mortgage. As the secondary market developed and geographical constraints on
mortgage origination were eliminated, they were able to take on more risk with the
knowledge that this risk could be offloaded in the secondary market. This has made
it possible for mortgage originators to offer many new kinds of mortgage products,
substantially reducing the remaining rigidities in the mortgage market.10 These developments and the lower interest rate environment were likely factors driving the
rise of the home-ownership rate from 64 percent in 1995 to 69 percent in 2005, after
being nearly flat for thirty years.11
How have all these changes affected the behavior of first-time home buyers? Table 2 describes how market outcomes of first-time home buyers have evolved over the
last 30 years. The table confirms that over this time it has become easier for first-time
home buyers to finance their first home purchase. Before and soon after mortgage
deregulation, there is essentially no change in the median income of a first-time house
buyer. In 1996, however, the median income of a first-time buyer is 7 percent lower
and by 2005 it is 15 percent lower. Despite having a lower level of income, these
buyers have been able to finance their house purchase with a much larger value to
income ratio. In 1976 the median house price is 1.5 time median income, in 1996 the
multiple is 2.5, and by 2005 it is 2.8. These more expensive houses are purchased with
a down-payment of just 11 percent of the house value on average in 2005, compared
9

Since the G.I. Bill of 1944, military veterans have been able to acquire Veterans Administration
mortgages without a down-payment.
10
Two concrete examples of the new mortgages available include those aimed at minimizing the
down-payment, so called “combo loans,” and those aimed at individuals with low creditworthiness,
so called “sub-prime” mortgages. Sub-prime lending has been around for a long time, but only
became widely available in the late 1990s. Recent turbulence in the sub-prime sector suggests that
this market is still developing.
11
Chambers et al. (2005) investigate the impact of changes in the structure of mortgages on the
home-ownership rate within a life-cycle model. Fisher and Quayyum (2006) estimate that changes
in the distribution of income and demographic factors account for roughly half the increase in the
home-ownership rate.

10

Table 2: Characteristics of First Time House Buyers
Statistic

1976

Median Real Income
$35,972
Median Price/Median Income
1.5
Mean Down-payment/Price
.18
Mean Monthly Payment/After-Tax Income
.23

1986

1996

2005

$35,481
1.9
.13
.30

$33,151
2.5
.12
.35

$30,281
2.8
.11
.40

Notes: The table entries are from various issues of The Guarantor, 1978-1998, the
2005 National Association of Realtors Profile of Home Buyers and Sellers, and 2005
American Housing Survey. The Real median income is based on the CPI. Mean
Monthly Payment/After-Tax Income before 2005 is from The The Guarantor. In 2005
we made an assumption about the average tax rate, .25, to calculate this variable.
to 13 percent in 1986 and 18 percent in 1976. To afford the higher value houses with
lower income, first time buyers pay almost double the share of their after-tax income
on mortgage servicing in 2005 compared to 1976. Overall, table 2 strongly suggests
it has become easier for first-time home buyers to obtain a mortgage.

2.5

The Baby Boom

Figure 3 plots variances of log per household residential investment over the previous
40 quarters (left-hand scale) and the fraction of household heads under the age of
35 (right-hand scale). This group of household heads has the highest propensity to
purchase a house for the first time. So this share variable can be viewed as a measure
of the relative number of potential first-time home buyers.
Housing volatility was relatively low in the early part of the sample, rises rapidly
to a peak in 1984 (corresponding to the period 1975-1984) and falls just as rapidly
after 1984 to new low levels by the mid-1990s. While housing investment volatility
clearly fell after 1984, there is a comparatively large housing recession in 1990-1991.
The most recent recession marks a dramatic change from previous experience as
residential investment hardly falls at all. As is confirmed by figure 2, the largest
amplitude fluctuations in residential investment are around 1972, 1978, 1982 and
1990. These episodes coincide with periods in which the share of potential first-time
11

Figure 3: The baby boom and the decline in volatility of residential investment
0.030

0.325

0.300
0.020

0.015

0.275

0.010
0.250
0.005

0.000

Share of Household Heads Under 35

Variance of Residential Investment

0.025

0.225
1959 1964 1969 1974 1979 1984 1989 1994 1999 2004

Note: Solid line – volatility of residential investment over the previous forty quarters
(left-hand scale); Dashed line – share of household heads under age 35 (right-handscale).
home buyers is at its highest levels. The drastically reduced fluctuations in the 1990s
and early 2000s occur as the share of potential first-time buyers reaches its lowest
post-WWII levels.
An increase in the share of potential first-time home buyers increases the bunching of households near the asset and income thresholds which historically households
have had to reach to obtain a mortgage. Other things equal, the degree of bunching
determines the magnitude of the response of first-time home buying to a given aggregate shock. This suggest the share of potential first-time home buyers in the economy

12

might have an important impact on the dynamics of residential investment.12
Two features of Figure 3 suggest that a credible explanation for the decline in
volatility cannot rely on just one of the structural changes we have emphasized. First,
the low share of under-35 year old household heads in the 1950s and 1960s is contemporaneous with relatively large fluctuations in residential investment. However, these
fluctuations do seem less violent compared to those in the 1970s and 1980s. Second,
the large drop in residential investment in the 1990 recession seems too far after the
elimination of regulation Q and the mortgage market deregulation which unleashed
growth in the secondary market and other innovations.13 However, many important
financial innovations occurred after this recession. So, even if the elimination of Regulation Q alone seems like it would have a hard time accounting for the decline in
housing volatility, mortgage innovation may still have played an important role.

3

Evidence on the Home Ownership Decision

The credibility of our quantitative analysis depends on us accurately representing both
the decision to own a home and any structural changes as they affect the market
for housing. As there have been structural changes in regulation, there have also
been secular changes in social norms regarding marriage and fertility. Given that
these social changes are intertwined with the decision to own a home, it is important
to understand what has remained the same and any secular changes in the home
ownership decision. This section describes three sets of findings along these lines.
First, throughout the sample the decision to purchase a home for the first time is
closely associated with marriage. Second, in the years immediately preceding an
individual’s first home purchase, income grows smoothly and savings accumulate
rapidly. Third, home ownership and marriage occur later in the life-cycle in the
second half of our sample sample compared to the first half.
12

See Jaffee et al. (1979) for an early investigation into the affects of the baby boom entering the
housing market.
13
The 1990 housing recession may have been exacerbated by a phenomenon similar in its affects
to Regulation Q. Around this time house prices in New England fell dramatically. This reduced
the capitalization of many mortgage lenders who at the time had to meet new regulatory capital
standards. Accordingly these banks were forced to curtail lending. Bernanke and Lown (1991)
discuss the influence of this ‘capital crunch’ on the 1990-1991 recession.

13

3.1

The Data

We study the behavior of individuals in the years before, during and after their first or
second purchase of a primary residence with the NLSY79 and PSID data sets. Both
data sets contain panel data on demographics, housing and income. Additionally, the
NLSY79 has panel data on wealth.14 The NLSY79 started in 1979 with a sample of
xxxx 14 to 22 year old individuals. These same individuals have been interviewed
each year until 2000, except in 1991 and every other year from 1995. The asset data in
NLSY79 begins in 1986. The PSID started in 1968 and has annual data on individuals
from the original sample and from individuals added to the sample population as they
enter the family histories of the original cohort. We use this data through to 1997.
To address oversampling of particular sub-populations in the surveys, our estimates
are based on suitably weighted observations.
We study unbalanced panels of individuals from the two surveys. In the NLSY
individuals are surveyed from an early age so we can almost always determine when
a first or later housing purchase is made. With the PSID we include individuals for
which the year of first purchase can be determined. We assume that anyone whose
first observation is as a renter and is no more than 25 years old has never owned.
In the PSID we exclude individuals who enter the sample over the age of 25. We
require an individual to be a head of household or spouse of head before they can be
classified as a homeowner.
The variables we focus on are housing tenure, income, net assets, marital status
and age. Individuals who are household heads or their spouses are both classified as
owning their house. All other individuals, including those living with their parents
or other family member, are classified as renters. An individual who gets married
to someone who already owns a house is classified as moving from renting to owning
in the year they get married. Income for an individual is defined as the sum of the
individual’s labor income and any spousal labor income. Net assets are defined, as
in Zagorsky (1999), as the sum of home value, cash, stockholdings, trust holdings,
business equity, car value, IRA holdings, certificates of deposit, 401(k) holdings and
non-car durables goods, less the sum of mortgage and other property debt, car debt
14

See Haurin et al. (1996) for an earlier investigation of net asset accumulation and the first home
purchase in the NLSY.

14

Figure 4: Home Ownership and Marriage Rates

.8
Percent
.4
.6
.2
0

0

.2

Percent
.4
.6

.8

1

B. Marriage Rate

1

A. Home Ownership Rate

18

20

22

24

26

28 30
Age

32

34

36

38

40

18

20

24

26

28 30
Age

32

34

36

38

40

38

40

.8
Percent
.4
.6
.2
0

0

.2

Percent
.4
.6

.8

1

D. Home Ownership Rate of Married

1

C. Home Ownership Rate of Unmarried

22

18

20

22

24

26

28 30
Age

32

34

36

38

40

18

20

22

24

26

28 30
Age

32

34

36

Note: These estimates are from the PSID. The ownership and marriage rates correspond to two 14-22 year old cohorts followed from 1968 (solid lines) and 1979 (dashed
lines).
and any other debt. Net assets are measured as the sum of individual and spousal
net assets.

3.2

Home Ownership and Marriage

In our theoretical model we assume that at an exogenously and randomly determined
stage during an individual’s life-cycle the individual’s level of housing subsistence
increases substantially. This motivates an increased demand for housing, which we
assume can only be accommodated by buying a larger home. We interpret this
“family” shock as the event of marriage. This sub-section examines evidence on
marriage and home ownership in the PSID to see if this interpretation is warranted.
15

Figure 4 illustrates home-ownership rates and marriage rates by age calculated
from the PSID. Panel A shows the evolving home ownership rates (percent of household heads who own their home) for two cohorts of 14-22 year-old individuals followed
from 1968 and 1979. The home ownership rate rises gradually as an individual ages
in both cohorts. For the later cohort, after age 22 or so, the home ownership rate
rise at a slower rate. In either sample, about three quarters of individuals who ever
by a house have already done so by age 35. We find that the marginal hazard of first
time home buying peaks at 27-29 for both cohorts.15 Panel B of shows that secular
changes in marriage rates are similar to those for home ownership. In particular, both
home ownership and marriage are occurring later in the life-cycle in the 1979 cohort
compared to the 1968 one.
Panels C and D show that marriage rates rise faster and remain higher for married
individuals compared to unmarried ones. About 20% of married individuals are home
owners in their early 20s. Unmarried individuals do not reach this rate of ownership
until their early 30s. There has been much less secular change in the home buying
behavior of unmarried individuals than in the overall marriage rate.
Figure 5 presents a formal characterization of the connection between home ownership and marriage. It plots conditional probabilities of being married in the years
surrounding a first or second home purchase. We want to assess the extent to which
marriage is associated with home purchase decisions. However, we know that age,
social trends and possibly the business cycle also influence the decision to get married.16 We want to abstract from these effects to isolate the marginal likelihood of
marriage at the time of the home purchase. This is accomplished by regressing a
dummy variable for whether the respondent is married on a set of dummy variables
for the years before, during and after the first or second purchase, plus year and age
dummies. The figure plots the fitted values and 95% confidence intervals for the year
relative to year of home purchase dummies. The omitted category for the “Years
15

This is based on a regression of a dummy variable for the first purchase on dummy variables
for age-bins and survey year. The coefficients on the age-bin dummies correspond to estimates of
first-purchase hazard rates relative to the omitted age-bin.
16
Age matters because in the early years of the life-cycle marriage rates are increasing with age.
Secular trends matter because social norms regarding marriage and childbirth have been changed.
The state of the economy could matter because, for example, the likelihood of being unemployed
might affect the decision to marry.

16

Figure 5: Marriage Near Home Purchases

.4
Probability
.2
0

0

.2

Probability

.4

.6

B. Secular Change in Marriage Rates

.6

A. Marriage Rates Near Home Purchases

−4

−3

−2 −1
0
1
2
Years from purchase

First Purchase

3

4

−4

Second Purchase

−2
0
2
Years from purchase
1968−1986

4

1979−1997

Note: These estimates are from the PSID. They correspond to coefficients from a
linear probability model with categorical explanatory variables. The explanatory
variables include dummy variables for each age, year and each number of years before,
during and after the first or second home purchase. The coefficients corresponding
to -4 to 4 years after the first or second purchase are plotted with 95% confidence
intervals.
from purchase” variable in the ‘First purchase’ model is individuals who never own.
The omitted category in the ‘Second purchase’ model is individuals who only ever
own one home in the sample.
Figure 5’s panel A reveals that marriage is more closely assoicated with the first
home purchase than the second (we get similar findings looking at the third purchase
as well). In the first purchase case respondents are approximately 45 percent more
likely to be married compared to respondents of the same age in the same year who
never buy a house. There is a big jump upward in the liklihood of being married
in the year of the first purchase. After the purchase there is little change in the
likelihood of being married. The qualitative pattern of marriage around the second
purchase is similar but much more muted. The peak is at 20 percent in the year of the

17

purchase, but there is not the jump at zero which occurs in the first case. We draw
two conclusions from this plot. First, marriage and home purchases occur at similar
times. Second, marriage is much more associated with the first than the second home
purchase.
Figure 5’s panel B illustrates the secular change in the dynamics of marriage and
the first home purchase. There is a tight connection between marriage and the first
home purchase in the 1968-1986 and the 1979-1997 sample periods. However, the
trend toward later marriages has weakened the relationship. Note that this reduction
is in addition to the secular decline in marriage rates by age which is accounted for
in our regression model by the year dummies. Since individuals purchase homes for
reasons other than family formation, we expect the association between marriage and
the first-purchase to decline if individuals marry later.

3.3

Income and Asset Accumulation

Figure 6 show household net assets and wage income around the time of the first
home purchase estimated from the NLSY. We use this evidence to assess our model’s
success in accounting for individual home buying behavior. The figure is constructed
using the same non-parametric regression as the one underlying Figure 5. The omitted
category for the “Years from purchase” variable is individuals who are never observed
owning. Income and net assets are displayed in proportion to the value at the time
of the purchase.
Figure 6 indicates that net assets and income rise together in the years -4 to -1. Net
assets double in the last year before the purchase. Income does to change in proportion
but still rises substantially between -1 and 0. The pattern of net asset accumulation
is consistent with survey evidence, reported in various issues of The Guarantor, that
households on average take about two years to accumulate their down payment. When
marriage occurs in our sample, the income and assets from the spouse are included in
the individual’s assets and income. Consequently some portion of the higher assets
and income in the figure is due to marriage. For example, in the year of the purchase
about half the rises in assets and income can be attributed to marriage. The sharp
increase in assets during the year of the home purchase reflects households’ equity in
the home and suggests that respondents in the NLSY typically pay a down-payment.
18

Figure 6: Household Net Assets and Income Near the First Home Purchase

Proportion of Value at Purchase
.5
1
0

0

Proportion of Value at Purchase
.5
1

1.5

B. Assets

1.5

A. Income

−4

−2
0
2
Years from Purchase

4

−4

−2
0
2
Years from Purchase

4

Note: These estimates are from the NLSY using the same regression model as in
Figure 5 with dependent variables net assets and labor income. The coefficients for
-4 to 4 years after the first purchase are plotted with 95% confidence intervals after
normalizing the coefficient at zero to one.
Overall, this figure is consistent with a respondent receiving a family shock which
leads her to accumulate rapidly to purchase a home. Even though the respondent
may expect to eventually purchase a home in her lifetime, accumulation does not
begin until she knows a home is needed soon. Income essentially level off after the
year of the purchase, but nets assets continue to grow so that they are about 25%
higher four years after the year of first purchase.

4

The Model Economy

In this section we describe our life-cycle model of housing. The evidence presented
in the previous section suggests that the event of buying a house for the first time is
closely related to family formation—individuals tend to become homeowners around

19

the time of marriage. We model this phenomenon as exogenous changes in housing
needs as individuals age through their life-cycle. When needs are high, individuals
desire a larger home which cannot be rented but must be purchased. The evidence
we have presented also suggests that historically households have faced financial constraints in order to become homeowners. We explicitly model two such constraints:
a down-payment constraint, which limits the loan-to-value ratio at the time of purchase, and a flow constraint which limits the interest payment on a mortgage to a
fraction of an individual’s current labor income at the time of purchase.

4.1

Individuals

Preferences The economy consists of a large number of ex-ante identical individuals who forever repeat the same “life cycle” of birth, work, retirement and death.
The transitions between the stages of life occur with fixed and known probabilities.
Individuals care about their future selves as much as they care about their current
self and so preferences are represented by
∞
X
Ut = Et
β j−t u(cj , hj − hj ), 0 < β < 1.
(1)
j=t

For the incarnation of the individual alive in period j, cj denotes the quantity of goods
consumed, hj denotes the quantity of housing services consumed, and hj denotes the
subsistence level of housing. Houses come in two sizes, hr < ho . We assume that the
larger size house cannot be rented. Tenure status at the beginning of the period is
denoted by x, where x = r corresponds to renting and x = o corresponds to owning.
For simplicity, below we drop time subscripts and use a prime symbol to denote next
period’s value of a variable.
Stages of the Life-cycle The state variable s controls both the life-cycle status
and labor earnings of individuals. Let s ∈ S = Y ∪F ∪R = {1, 2, . . . , N }∪{N +1, N +
2, . . . , 2N } ∪ {2N + 1, 2N + 2, . . . , 3N }. Individuals go through three stages of life.
When s ∈ Y, an individual is a young worker with low housing needs, h = 0) When
s ∈ F , an individual is a family worker with high housing needs, h = hF . Finally,
when an individual’s state transits to s ∈ R, the individual retires and housing needs
decrease to h = 0. We assume hF < hr .
20

Non-retired individuals supply one unit of labor inelastically and face idiosyncratic
uncertainty with respect to their labor productivity. An individual in state s ∈ Y ∪ F
is endowed with e(s) efficiency units of labor, each unit being paid after-tax wage
rate w = (1 − τ )ŵ, where τ is a labor income tax and ŵ is the before-tax wage rate.
The revenues from the labor income tax are used to operate a pay-as-you-go social
security system. All retired individuals are entitled to a social security payment equal
to a fraction, θ, of average before-tax earnings of the working population. To keep the
notation consistent with working individuals, we let e(s) = θe/(1−τ ) if s ∈ R, where
e is the average labor productivity of the working-age population. Given the simple
structure of this social security system, it can easily be shown that τ = θµR /(1 − µR ),
where µR is the fraction of the population that is retired.
The process governing an individual’s
matrix Π,

ΠYY

Π =  0N
ΠRY

state over time is described by the Markov

ΠYF 0N

ΠFF ΠF R  ,
0N ΠRR

where 0N denotes an N × N matrix of zeros and the other terms are non-zero N × N
matrices. Since individuals need to go through an entire life-cycle, the probability of
going from set Y to set R is zero. Similarly, the probabilities of transiting from set F
to set Y and set R to set F are also zero. The elements of matrix ΠYY control productivity levels while an individual is a young worker, and those of matrix ΠF F control
productivity levels of a family worker. The matrices ΠF R and ΠRR are diagonal. The
matrix ΠRY controls the probability of dying and the magnitude of intergenerational
income persistence. We use πss0 to denote individual elements of Π.
At the same time as death, a new generation of individuals of size (1 + g) are
born. The productivity levels of the newborn are controlled by the elements of the
matrix ΠRY as follows:


θ1 (1 − δ1 ) · · · θN (1 − δ1 )


..
,
ΠRY = 
.


θ1 (1 − δN ) · · · θN (1 − δN )
where δn is the probability of remaining retired if you retired as a type-n individual,
and [θ1 , . . . , θN ] is the part of the invariant distribution Π associated with the young
21

stage of life. Notice that as written, the matrix ΠRY assumes that there is no intergenerational income persistence if the probability of dying is the same no matter which
type of individual you were when you retired (δj = δ for all j). This is because each
individual has the same probability of being any of the N types of young individuals,
regardless of the parent’s type at the time of death.
Borrowing Constraints Individuals accumulate wealth with two types of assets:
owner-occupied houses, ho , and a generic asset called deposits, d, which pays interest id . Let a denote current net worth. Renters face a non-negative savings restriction,
a0 ≥ 0. Homeowners may borrow against their house, i.e. they can acquire a mortgage. Borrowing against a home involves two constraints which affect the size of the
down-payment and the amount of the mortgage payment relative to an individual’s
income. These constraints only apply when an individuals is changing from being a
renter to a homeowner. Once the individual has a mortgage, the constraints no longer
apply.
The downpayment constraint says that a mortgage acquired in the current period,
m , is limited to be no more than a fraction γ1 of the value of the home so that
m0 ≤ (1 − γ1 )ho . Current savings of an individual who chooses to be a homeowner
next period are a0 = d0 + ho − m0 . It follows that savings must be at least as big as
the minimum down-payment on the house: a0 ≥ γ1 ho .
0

The flow constraint says that the loan payment must not exceed a fraction, γ2 , of
labor income in the current period. Letting i0m denote the interest rate on a mortgage
contracted today, the flow constraint is i0m m0 ≤ γ2 we(s). We assume im ≥ id so that
individuals at least weakly prefer paying their mortgage before accumulating other
assets.17 Consequently, d0 = 0 when the flow constraint binds and we can write the
flow constraint as a0 ≥ ho − γ2 we(s)/i0m .
Comparing the two constraints reveals that there exists a threshold value of e(s)
such that both constraints hold with equality. For all values of e(s) exceeding this
threshold, the flow constraint is irrelevant, and for all values below it, the downpayment constraint is irrelevant. Therefore, we can combine the two constraints into
17

When the interest rates are identical we assume individuals act as if the mortgage rate is higher.

22

a single constraint:
a0 ≥ γ(s, x, x0 )hx0 ,
where

(
γ(s, x, x0 ) =

(2)

0,
if x0 = r or x = x0 = o.
max {γ1 , 1 − γ2 we(s)/(i0m ho )} , if x = r and x0 = o.

Recursive Formulation of an Individual’s Problem The problem faced by an
individual is to choose a sequences of consumption, asset holdings, and housing tenure
choices to maximize (1), subject to (2) and the budget constraint
c + ph hx + a0 = we(s) + R(a, x)a + T

(3)

holding in each period. Here ph denotes the rental price of housing and T is a lump
sum transfer described below.18 The function R(a, x) in (3) is given by
(
1 + id ,
if x = r or if x = o and a ≥ ho ;
R(a, x) =
1 + im − (im − id )ho /a, if x = o and a ≤ ho .
This last expression takes advantage of two key facts. First, since im ≥ id , individuals
weakly prefer renting to owning a house of size hr . Second, since individuals always
at least weakly prefer paying their mortgage to accumulating non-housing assets, no
individual with a fraction of equity in [γ(s), 1) has strictly positive non-housing assets.
Only renters and individuals who own their house outright accumulate non-housing
assets. It follows that we can infer the size of an individual’s mortgage or deposits
from their savings and their tenure choice.
To describe the recursive formulation of the individual’s problem, denote the value
function of an individual in earnings state s, with net worth at the beginning of the
period a, and tenure status x, by W (s, a, x). The value function depends on the
individual’s tenure choice for tomorrow.19 Specifically,
n
o
W (s, a, x) ≡ max Vr (s, a, x) , Vo (s, a, x) ,
(4)
18

The rental price of housing appears in the budget constraint for a current homeowner as an
implicit payment. To keep the notation simple in writing this budget constraint we have ignored the
fact that assets at the beginning of the period may have been subject to the estate tax. However
this budget constraint is valid for the recursive formulation of the individual’s problem.
19
This value function and the ones described below also depend on the distribution of individuals
over the state variables. We suppress this term for simplicity.

23

where the first term is the value of choosing to rent tomorrow and the second term
is the value of choosing to own tomorrow.
For x0 ∈ {r, o}, these value functions are defined as follows. The value function of
an individual with state variables (s, a, x) for s ∈ {Y ∪ F } is given by
Vx0 (s, a, x) =

max0

a0 ≥γ(s,x,x )hx0 , c

X
¡
¢
πss0 W (s0 , a0 , x0 )
u c, hx − hs + β

(5)

s0 ∈S

subject to (3). The value function of an individual with state variables (s, a, x) for
s ∈ R is given by
Vx0 (s, a, x) =
max

a0 ≥γ(s,x,x0 )hx0 , c

¡

¢

u c, hx − hs + β

"

X

πss0 W (s0 , a0 , x0 ) +

s0 ∈R

X

#
πss0 W (s0 , (1 − τE )a0 , r)

s0 ∈Y

(6)
subject to (3). Here τE is an estate tax paid by newly born individuals. Note that if
τE = 1, then individuals are born with zero non-housing assets. Proceeds from estate
taxes are redistributed evenly among living individuals as the lump-sum transfer, T .
Individuals are always born without a house. Upon death, houses are liquidated and
put on the market for sale. Newborn young individuals are born without a house and
so must rent.

4.2

Producers

Firms maximize profits
f (k, l) − wl − pk k,
where f (k, l) is a constant returns production function, k denotes non-residential
capital used in production, l denotes the quantity of labor employed, measured in
efficiency units, and pk denotes the rental price of non-residential capital. We assume
that producers’ output can be costlessly transformed into consumption goods, and
new residential and non-residential capital. Consequently, the prices of these goods
are all equal to one in a competitive equilibrium.

24

4.3

Financial Intermediaries

Capital investment and mortgage lending are undertaken by over-lapping generations
of two-period-lived financial intermediaries. In the first period each intermediary
issues one period securities called deposits, d0 , which pay interest i0d . The funds raised
in this way are invested in capital and mortgages. Investing in residential capital
involves purchasing the capital, h0 , at the end of the first period, renting it in the
following period at price p0h , and selling the undepreciated portion at the end of the
second period (to newly born intermediaries). Investing in non-residential capital, k 0 ,
is the same except in this case the rental price is p0k . Investing in mortgages involves
lending m0 at the interest rate i0m . The mortgages are paid in full with interest at
the end of the second period. To introduce a spread between borrowing and lending
rates of interest faced by individuals, we assume there is a proportional cost to issuing
mortgages. Specifically, intermediaries face the resource constraint
(1 + φ)m0 + h0 + k 0 ≤ d0 ,

(7)

where φ ≥ 0 is the cost of mortgage lending, in units of the output good. Intermediaries behave competitively and maximize profits subject to (7). The necessary
conditions for optimality imply
ph = im + δh ,

pk = id + δk ,

and im = id (1 + φ),

(8)

where δh and δk denote the depreciation rates of residential and non-residential capital respectively.20 Notice that the spread between the mortgage and deposit rates
depends on the level of the deposit rate.

4.4

Stationary Competitive Equilibrium

A stationary competitive equilibrium consists of a value function W (s, a, x), decision
rules for savings g(s, a, x) and for tenure choice h(s, a, x), prices {id , im , w, ph , pk }, a
fiscal policy {τ, θ, T, τE } and a measure λ(s, a, x) such that
20

We interpret depreciation of housing as follows. Each house is subject to a probability, δh , of
burning down and all owners of houses participate in an actuarially fair insurance scheme. Houses
are replaced instantaneously.

25

1. The value functions and associated policy rules solve the recursive individual
problem as given by equations (4)–(6);
2. Factors are paid their marginal products: pk = f1 (K, L), w = f2 (K, L) and (8),
where K and L denote aggregate non-residential capital and labor input;
3. Aggregates are consistent with individual behavior, that is λ(s, a, x) is generated
by:
λ(s0 , a0 , x0 ) =
( P
R
P
0
0
x∈{r,o} a∈A(s0 ,a0 ,x0 ) λ(s, da, x), unless s ∈ R, s ∈ Y, x = o
s∈S πss0
0,

otherwise

where
( ©
ª
0
0
a|g(a,
s,
x)
≤
a
,
h(s,
a,
x)
=
x
, unless s ∈ R, s0 ∈ Y
©
ª
A(s0 , a0 , x0 ) =
a|g(a, s, x) ≤ (1 − τE )a0 ,
otherwise;
4. The social security (given θ) and estate tax (given τE ) systems are individually
self-financed: τ = θµR /(1 − µR ), and
X
X Z
πss0
g(s, a, x)λ(s, da, x);
T =
s∈R, s0 ∈Y

x∈{r,o}

5. Markets clear:
K +H =

X X Z

g(s, a, x)λ(s, da, x),

s∈S x∈{r,o}

where H is the aggregate stock of housing. The goods market must clear by
Walras’ law.

5

Calibration

We compare the impact of aggregate shocks on residential investment under a post1984 benchmark calibration to the impact under a calibration which represents the

26

pre-1984 period. This section describes how we assign values to the model’s parameters in the two calibrations. We use the post-1984 period as our benchmark because
this is where the micro evidence on income and asset accumulation is from. The
structural changes we consider are modelled by changing key model parameters from
their values in the benchmark calibration.

5.1

Post-1984 Benchmark Calibration

The functional form of the utility function is given by
u(c, h − h̄) = ln[(1 − η)cρ + η(h − h̄)ρ ]/ρ,

η ∈ [0, 1], ρ ≤ 1.

We need values for parameters governing preferences, {β, η, ρhF }, the income process,
{Π, e, θ, τ, τE }, the production technology, {α, δh , δk }, the size of homes, {hr , ho }, and
mortgage constraints, γ. Our strategy is use direct evidence to assign values to the
technology, income and mortgage parameters and then to choose the preference and
house size parameters to bring the model as close as possible to several empirical
statistics. Table 3 summarizes the parameters values we use.
The production function is f (k, l) = k α l1−α and the technology parameters are
α = 0.3, δk = 0.092 and δh = 0.014. These values correspond to means of labor’s
share and depreciation rates for nonresidential and residential capital calculated from
NIPA data for the period 1947-2005. Our measure of nonresidential capital includes
the stocks of private nonresidential capital and consumer durables. Residential capital
corresponds to the stock of private residential capital.
For the income process we assume N = 3 so that there are three income states in
each life-cycle stage. Two objects need to be specified for each stage: the expected
duration of the stage and the process for income conditional on being in a given stage.
We assume that life begins at 18. The duration of the first stage of life is chosen so
that the fraction of under-27 individuals in the economy that have not transited to
the second stage equals the fraction of under-27 individuals who are unmarried for
the cohort born in the period 1958-1967 reported in Table IV of Caucutt et al. (2002).
We assume the second stage duration averages 37 years and the third 9 years. The
conditional income process for the young and family stages are constructed using
the bias-corrected version of the Tauchen and Hussey (1991) algorithm proposed by
27

Floden (2007). We select autocorrelations and innovation variances for the young
and family income processes using the life-cycle income process estimated from the
PSID by Storesletten et al. (2004), reported in their table 2 and figure 1. The data
underlying these estimates spans the two sub-samples which are the focus of this
study. We assume that the life-cycle conditional variances they report are a weighted
sum of variances from the early and the later sub-periods, and that the later variance
is 36 percent higher than the early variance. This percentage increase is taken from
Heathcote et al. (2004) who report that the cross-sectional variance of wages for the
20-59 age group is this much larger in 1990 compared to 1970. Both autocorrelations
are set to .95. To complete the specification of income for the young and family stages
we normalize average efficiency units of work to unity and assume average income for
a family individual is 1.4 times that of a young individual. The latter value is taken
from Castañeda et al. (2003). The replacement ratio for retirees is θ = 0.4, which is
taken from Mitchell and Phillips (2006). We set the labor tax, τ, to the value which
finances the social security system. The estate tax is τE = 1 so that individuals are
born without any assets.
The last four parameters we specify directly are the discount factor, the downpayment constraint, the size of owned houses related to rented houses, and the population growth rate. In keeping with related studies, we fix β to 0.97. The down-payment
constraint, γ is set to .12. This is value of the mean down-payment to income ratio
for 1996 in Table 2. We specify the size of owned relative to rented homes to 2.5. This
is based on an analysis of the American Housing Survey described in the appendix.
The population growth rate is set equal to the average growth rate of the number of
households from 1980-2005, 1.3 percent.
The parameters, η, hF and hr , are chosen to minimize the sum of squared deviations from their empirical counterparts of the model’s predictions for the ratio
of nominal spending on housing to the sum of nominal consumption and housing
spending, and the home-ownership rates for 20–24, 20–27, 20-30 25–64, 28–64 and
31-55 year old household heads. The first statistic is based on NIPA data for the
period 1947–2005, and the remaining six statistics are averages from the 1990 and
2000 censuses.
The last parameter to select is ρ. We use this parameter to match the trajectory

28

Table 3: Post-1984 Benchmark Calibration
Parameter
β
η
hr
ho
h̄Y
h̄F
h̄R
γ
α
δk
δh
θ

Description
Discount factor
Weight on housing preferences
Size of rental units
Size of owner-occupied houses
Minimum house size for young
Minimum house size for family
Minimum house size for retired
Minimum down-payment
Capital share of income
Rate of depreciation of capital
Rate of depreciation of housing capital
Social security replacement ratio )

Value
0.97
0.18
1.65
4.125
0.00
1.53
0.00
0.12
0.30
0.096
0.015
0.40

of net assets after a first home purchase shown in Figure 6. The parameter ρ controls
the degree of complementarity between housing services and consumption. If ρ = 1
then housing and consumption are perfect substitutes. As ρ is reduced, the goods
become more complementary and in the limiting case ρ → −∞ the goods are perfect
complements, i.e. preferences are Leontieff in consumption and housing. This latter
extreme case is useful for understanding how ρ affects savings. In the Leontieff case
consumption is chosen to be a fixed proportion to housing, reducing the incentive to
save after becoming a home-owner compared to higher values of ρ.

5.2

Pre-1984 Calibration

We consider four differences in the environment facing individuals before 1984 compared to after: a tighter down-payment constraint, faster household growth, earlier
marriage and a lower cross-sectional variance of earnings. The borrowing constraint
is set to γ = .18, the value of the mean down-payment to income ratio for 1976 reported in Table 2 The effect of the baby-boom on home-buying is captured by raising
the population growth rate to 2.0 percent, which is the average household growth

29

Table 4: Statistics for the Benchmark Economy
Pre-1984
Data
Model

Post-1984
Data
Model

Aggregate Quantities
I/Y
K/(K + H)
ph H/(C + ph H)

0.26
0.56
0.20

0.32
0.42
0.23

0.25
0.56
0.21

0.29
0.43
0.21

Home-ownership Rates
20–24
20–27
20–30
25–64
28–64
31–55

0.24
0.31
0.38
0.68
0.71
0.72

0.25
0.33
0.38
0.70
0.74
0.76

0.17
0.24
0.32
0.65
0.68
0.71

0.21
0.28
0.33
0.65
0.68
0.69

rate from 1955-1979.21 Earlier marriage is modelled by accelerating the average wait
to transit from being a young to being a family individual by two years. Doing this
matches the fraction of under-27 who are unmarried for the birth cohort 1948-1957 reported by Caucutt et al. (2002). We decrease the cross-sectional variance of earnings
by the 36 percent value described above.

5.3

Characteristics of the Pre- and Post-1984 Calibrations

Table 4 displays some key features of the stationary equilibrium associated with each
calibration. It shows that our model does well replicating US data in the two sample
periods. In particular, the model does well reproducing the life-cycle profiles of homeownership rates in the two periods, and the fact that even though the down-payment
constraint is less stringent in the Post-1984 calibration, the home-ownership rates are
lower than for the Pre-1984 case. The model is less successful in its predictions for
the allocation of capital across non-residential and residential uses; it predicts too
21

The source is the Census Bureau. We end the sample in 1979 since the data after 1980 are
calibrated to the 1980 census which introduces a break in the level of the household series in that
year. In addition, there appears to be a clear break in the growth rate of households in 1980.

30

Figure 7: Net Assets and Income in the Model and Data

Proportion of Value at Purchase
.5
1

1.5

B. Assets

0

0

Proportion of Value at Purchase
.5
1

1.5

A. Income

-4

-2
0
2
Years from Purchase
Data

4

-4

Model

-2
0
2
Years from Purchase
Data

4

Model

Note: Solid line – Net assets; Dashed Line – Labor income. Net assets and income
have been normalized by the value at purchase.
much housing compared to the data. However, the amount of spending on housing is
consistent with the data.
Figure 7 compares the profiles of income and assets around the first home purchase
in our model to our estimates discussed in section 3.3. This figure shows that our
model is consistent with the basic patterns in the data: both income and assets rise
leading up to purchase, income flattens out afterwards, and wealth continues to grow.
The main drawbacks of the model are that income and assets grow too slowly and
the level of assets is too high in the four years preceding the purchase. Still, we think
the qualitative success of the model justifies using it to study aggregate dynamics.

31

6

Aggregate Dynamics

We investigate the dynamics of residential investment after two different kinds of
aggregate shocks, an “interest rate” shock and an “income” shock. We calculate
the investment dynamics starting from two stationary equilibria corresponding to
our benchmark and pre-1984 calibrations, under the prevailing regulatory and demographic regimes.

6.1

Methodology

For our dynamic analysis we assume the economy is small relative to world financial markets and all goods are tradeable, but labor is not tradeable. Under our
assumptions the real interest rate is exogenous, but all other prices relevant to the
individual’s problem are determined endogenously. We study two kinds of transitory
disturbances, to the real interest rate and to labor augmenting technical change.
Consider the interest rate shock; the income shock case is similar. For simplicity
we suppose there are no mortgage transactions costs, φ = 0, so the deposit and
mortgage rates both equal the world interest rate. Suppose at date 0 the economy
is in stationary equilibrium where the interest rate is i∗ and the current account is
balanced. From date 0 we suppose the real interest follows a path lasting T years
given by i1 , i2 , ..., iT , where it denotes the rate of interest from year t − 1 to t. After
the T ’th year the interest rate reverts to i∗ and the economy converges to the initial
stationary equilibrium with a balanced current account.
We use backward induction to solve for the equilibrium path. We solve for the
date t prices pk,t , ph,t and wt as well as the date t stock of domestic nonresidential
capital, Kt , using the following four equations
pk,t = it + δk

(9)

ph,t = it + δh

(10)

pk,t = F1 (Kt , L)

(11)

wt = F2 (Kt , L)

(12)

Equation (9) is the international arbitrage condition, equation (10) is the domestic
capital arbitrage condition, and equations (11) and (12) are the efficiency conditions
32

from producers’ capital and labor input decisions. Equations (9)-(12) also hold after
date T so in date T + 1 and after all prices faced by individuals correspond to the
stationary equilibrium prices. Therefore the value function of an individual for date
T + 1 is the stationary equilibrium value function, W ∗ (s, a, x). Using this fact we can
solve backwards for the date T, T −1, ..., 0 decisions of individuals and value functions
of individuals using equations (4)-(6). Once we have individual decisions for dates
0, 1, 2, ..., T the transition equation for the distribution of individuals over the state
space is used to compute the distribution of individuals over the state space for dates
0, 1, 2, ..., T, ... These distributions can in turn be used to compute aggregate demand
for residential capital, HtD .22 Our interest is in the dynamics of residential investment
in response to the interest rate shock. This is given by
D
D
Xh,t = Ht+1
− (1 − δh )Ht+1
.

This approach to studying aggregate dynamics is analogous to examining an impulse response function, which is the traditional approach to dynamic analysis in the
business cycle literature. While our approach is analogous to traditional exercises
it is different in two key respects and these are worth discussing. First, we do not
compute the closed economy transition path. That is, the interest rate is taken to
be exogenous, whereas it is typically endogenous in the business cycle literature. We
have taken this approach so we are able to approximate how the economy responds
to something resembling a monetary policy shock, without having to model such a
shock explicitly. It is important to consider the impact of monetary policy shocks
since they are widely viewed as having a significant impact on residential investment.
Second, we have not solved the model taking into account the aggregate uncertainty
we are implicitly assuming the agents in our model face. We do this to avoid taking a
stand on what constitutes the entire set of shocks affecting the economy. The shocks
we consider are meant to be a stand-in for all shocks which affect the economy.

33

Figure 8: Responses to Interest Rate Shock
1.1
Early period
Early period with Reg−Q
Later period
1.05

1

0.95

0.9

0.85

6.2

0

1

2

3

4

5

6

7

8

Interest Rate Shocks

Our interest rate shock consists of an initial increase in the interest rate followed
by a gradual decline.23 This pattern is consistent with estimates of the response of
interest rates to a monetary policy shock in the literature. Figure 8 displays three
paths of residential investment in the face of the temporary deviation of the interest
rate from its stationary equilibrium level. This shows the dynamics implied by the
pre-1984 calibration with and without Regulation Q and the post-1984 calibration.
Regulation Q is modelled as a tightening of the down-payment constraint from its
22

The path of the current account is solved for using the goods market clearing condition. The
corresponding path of the capital account is consistent with the dynamics of Kt and the total demand
for domestic residential capital. We do not compute the paths of the current and capital accounts
since we do not need them to solve for the residential investment dynamics.
23
Specifically, the interest rate deviates from its stationary equilibrium level for eight periods. The
specific magnitudes of deviation are .0035, .0021, .0011, .0007, .0005, .0004, .0003, .0002

34

Table 5: Quantifying the Effects of Structural Change
Variance
Ratio
.21

US Data
Interest rate shock
Total (including Reg Q)
Without Reg Q
Only change down-payment constraint
Only change g
Only change σy
Only change marriage delay
Only change g, σy and marriage delay
Income shock
Total
Only change
Only change
Only change
Only change
Only change

down-payment constraint
g
σy
marriage delay
g, σy and marriage delay

.06
.12
1.13
.03
.23
.31
.11

.03
.62
.03
.54
.32
.03

calibrated value, .18, for two periods: .22 in the first period of the experiment and
.2 in the second period. The figure illustrates two facts. First, there is a dramatic
dampening in the response of residential investment in the post-1984 case compared
to the two pre-1984 cases. Second, comparing the two pre-1984 cases we see that
Regulation Q can only be part of the story of the decline.
We use a statistic analogous to the variance ratio reported in Table 1 to quantify
the impact of structural change. Specifically, we calculate the sum of squared responses minus one for eight periods for each response path we consider and construct
ratios of responses representing the later period to responses representing the early period. If the aggregate dynamics are approximately linear in the domain of the shocks
then these ratios should provide an accurate representation of how volatility changes
between the two periods. Table 5 displays ratios for the two shocks corresponding to
35

various scenarios designed to quantify the impact of the different structural changes.
The interest rate experiments are described in the middle panel of Table 5. For
these shocks the table indicates that the overall impact of all the structural changes is
to reduce volatility to just 6 percent of what it was in the Pre-1984 case. This seems
very large, and is even larger than the decline in the data. Even without Regulation
Q in the Pre-1984 period, the decline is to 12 percent. Comparing these cases we see
that the elimination of Regulation Q would drop volatility by just 50 percent. So most
of the effects must be due to the other structural changes. The remaining interest rate
experiments are designed to disentangle the impact of these other structural changes.
In each of these cases we recompute the stationary equilibrium of the Pre-1984 period
with only the structural changes indicated in the table, and use this as the Post-1984
case when forming the variance ratio.24 What these experiments show is that the drop
in the variance ratio including all the structural changes is entirely due to changes
in population growth, increased marriage delay and the change in the cross-sectional
variance of income.
What underlies this finding? First note that variation in residential investment
in the model is driven entirely by changes in the numbers of individuals choosing to
own, since the magnitude of the difference in house sizes is fixed. Now, consider the
household’s decision to own a home. This is determined by a threshold rule. For
given income and tenure status, there is a threshold level of wealth which determines
whether the household will own in the next period. Wealth below the threshold means
the household will rent, otherwise it will own. After a shock, the thresholds change
and the impact on residential investment depends on how many individuals change
their tenure choice compared to the previous period. This in turn depends on the
distribution of wealth near the threshold. It turns out that almost all the individuals
who change there tenure decision are in the low and middle income state of the family
stage of life, and of these, almost all the difference between the Pre- and Post-1984
scenarios comes from those individuals who before the shock would have switched
from renting to owning, but after the shock choose to stay renting. So, by analyzing
the distribution of wealth around the threshold for these individuals we can find the
source of the effects in our model.
24

This is not an orthogonal decomposition of variance so there is no reason for the effects to add
up to the total.

36

What we find is that each of the three non-housing-specific structural changes
has, for different reasons, a large impact on the number of individuals affected by
the shock. The drop in population growth means that there is a smaller fraction of
individuals in the family stage of life in the later period. Remarkably, as the table
indicates, this alone could be the source of the volatility decline. The increase in
marriage delay means that when individuals reach the family stage they are wealthier.
This has the effect of shifting the distribution of wealth to the right, lowering the
fraction of individuals near the threshold. A similar affect occurs with the change in
cross-sectional variance of income. In this case, increased savings occur because of the
induced increase in the precautionary savings motive. The table suggests these effects
are smaller than with the decline in population growth, but still sizeable. Finally,
notice that the reduction in the down-payment constraint actually leads to an increase
in volatility.

6.3

Income Shocks

The income shock is modelled as a proportional shift in the life-cycle profile of labor
productivity. As such it is similar to a technology shock in a real business cycle
model, except that the interest rate is kept at its stationary equilibrium level. The
quantitative findings for the income shock experiments are in the bottom panel of
Table 5. The findings here are very similar to the interest rate shock experiments. In
particular, the overall impact on volatility is large and mostly due to the non-housingspecific structural changes. Of these, the reduction in population growth continues to
be the most important driver of the volatility decline. The only substantive difference
with the interest rate shock case is that here the reduction in the down-payment
constraint does lead to a modest decline in volatility.

7

Conclusion

Our findings strongly suggest that even in the absence of structural change in the
housing market, we should have seen a large decline in the volatility of residential
investment. While our findings are based on studying just two kinds of aggregate

37

disturbance, we think the findings hold more generally. This is because the principle
mechanism driving the findings, changes in the distribution of wealth relative to the
threshold determining housing tenure choice, should influence the response of the
economy to any kind of shock in a similar way. Therefore, we think our findings
provide support for the view that the “Great Moderation” in aggregate fluctuations
is at least in part driven by structural change and is not just a result of the economy
being subject to smaller aggregate shocks.

Data Appendix
1. Calculating Housing Tenure Transitions in the PSID
The variables in the dataset are obtained from the PSID 1968-1996 family and
individual core sample files. An agent is selected into the sample if she lives in the
household unit at the time of the interview, and if either one of two criteria hold: i)
the agent is a household head or spouse ii) the agent is not a head or spouse but has
age 18 or greater
The household weight variables are adjusted to match the household age distribution from the Census Current Population Survey (CPS). For each age bracket in
the CPS, we divide the Census-determined number of households by the PSID family weight determined number of households to obtain the scale factor by which the
PSID weights over- or under-represent an age group. To obtain Figure 1, the PSID adjusted weights from above are partitioned to avoid double-counting when a household
changes residence. For example, when individuals from separate households marry to
form a single household, it is desirable to avoid counting the home purchase twice.
Thus, individuals are assigned a portion of the family weight, but rather than dividing
this weight by the number of household members, we use the PSID individual weights
to determine an individual’s relative importance within the household, i.e. [(individual weight)/(sum of household unit’s individual weights)]*(family weight) determines
the individual’s contribution to the family weight. The sum of these contributions
gives back the family unit’s weight.
38

The sample for the Rent-to-Own series consists of individuals who are observed
transitioning from renting to owning to obtain more housing. Specifically, if the home
purchased at time t is physically larger than the one rented in time t − 1. The house
size is measured by the number-of-rooms indicator from the PSID family file. We
further restrict the sample to those individuals who respond to the question “Why
did you move?” with the by either responding for “Productive” or “Consumption”
reasons. We chose these cases because they reflect deliberate moves to obtain more
housing. The resulting probability is calculated using the weighting procedure from
above restricted on the population who have non-missing own/rent observations at
time t and time t-1. Similarly, the Own-to-Own series identifies individuals who are
observed owning in both t and t-1, but who have a move recorded in the time since
the last wave’s interview.
A first time home buyer is defined as a head or spouse who satisfies the following:
i) observed entering the sample at age 25 or less ii) observed entering the sample
renting iii) has non-missing rent observations until first instance of homeownership.
Anyone transitioning from renting to owning or owning to owning (with a recorded
move in residence) who is not buying their first home is “repeat buying.” A nonhead, non-spouse family unit member is automatically assumed to be renting. An
individual is classified as never owning if their tenure choice is observed in every year
they are in the sample, and they are in the sample for at least fifteen years.
2. Net Assets and Income in the NLSY
Net assets and income include those for the survey respondent and their spouse,
if the spouse lives with the respondent. Income is just labor income and net assets
includes assets from all sources listed in the survey. To be a classified a first-time
home buyer we must observe the tenure choice of the respondent for all periods up to
the first time they are recorded as owning. To classified as never owning, the tenure
choice of the respondent must be observed in all years they are in the sample up until
the last year the survey is annual. We use the weights provided by the NLYS in our
estimation to correct for the oversampling of the poor and members of the military.
3. Calculating Relative Size of Houses for Owners and Renters in the American
Housing Survey
From 1985 The American Housing Survey (AHS) obtains the square feet size of
each home in the survey. We use this measure to estimate the relative size of owned
to rented homes in a manner as consistent as possible with out model. This measure
of home size has the drawback that it does not account for ameneties. It has the
advantage that it excludes land, which would not be the case if we focused on home
value and rent. This is relevant since residential investment only measures spending
39

Table 6: Average Home Size of Owned Versus Rented Homes
Measure 1985
(1)
(2)
(3)

2.5
2.4
2.8

1989

1993

1997

2001

2005

2.4
2.3
2.7

2.4
2.4
2.7

2.1
2.1
2.6

2.7
2.6
3.1

2.4
2.4
2.7

on residential structures.
We consider three different measures of home size. Each involves computing average size of owned versus rental units, but calculates size using different approaches.
Measure (1) focuses on rental homes with only one person in household versus owned
homes with more than one person in the household. In addition this measure restricts the sample to household heads under 30 for rental units and household heads
in the age range 30-55 for owned units. (2) Restrict sample to households without
children and unmarried for rental units versus owned units with married households.
In addition space is measured as per adult for renters (except opposite sex couple
households) versus no per adult correction for owners. (3) Measure (2) with the age
restrictions of (1).
These alternative measures are intended to get closer to the concept we have in
the model: there are two kinds of space - one you live in before the family shock and
one you live in after the shock. By not splitting out the data like in (1)-(3), say by
only looking at rental and owned units without any other restrictions to the sample,
we mix up people who rent but have had the family shock with renters who have
yet to have the shock. In the data one can rent a larger unit but not in our model.
Measures (2) and (3) get at the idea that before the family shock, if you have a roommate the whole space in the housing unit is not really your own (at the very least you
need your own room and you probably share the rent). Measures (2) and (3) do not
account for homosexual couples. Not using per adult space with married cases is an
attempt to get at the idea that other things the same two unmarried individuals need
more space to be as well off as a married couple. The advantage of (1) is that there is
no need to make implicit assumptions about how space for two or more people maps
into “virtual” space for one. Opposite sex couples are assumed to be like married
couples. At the very least they probably do not need an extra bedroom.
Table 6 summarizes our findings for selected years. The numbers refer to the
measures listed above. Note that square feet is topcoded in the AHS. We leave the
top coded values in for our calculations but expect that the ratios are downward
biased because more owned dwellings tend to be topcoded than rented dwellings.
With the exception of 1997 (where there are much fewer observations) the ratios
40

are remarkably stable. This is true even though the size of the units varies a fair
amount and is growing over time. Excluding 1997, these ratios range between 2.3
and 3.1. Excluding the largest and smallest ratio leaves a range between 2.4 and 2.8,
suggesting 2.5 is a “conservative” choice. Note that these ratios are calculated for
the same period we have asset data from the NLSY so they are directly relevant to
our benchmark calibration period. Since the ratios seem to be relatively stable even
as the average size of a type of unit is generally growing over time, we see no reason
to choose a different house size ratio for the pre-1984 sample.

41

References
Ahmen, S., A. Levin, and B. Wilson (2001). Recent U.S. macroeconomic stability:
Good luck, good policies, or good practices? Unpublished manuscript.
Bernanke, B. S. and C. S. Lown (1991). The credit crunch. Brookings Papers on
Economic Activity (2), 205–239.
Campbell, J. R. and Z. Hercowitz (2004). The role of households’ collateralized debt
in macroeconomic stabilization. Unpublished manuscript.
Castañeda, A., J. Dı́as-Gı́meńez, and J.-V. Rı́os-Rull (2003). Accounting for earnings
and wealth inequality. Journal of Political Economy 111 (4), 818–857.
Caucutt, E., N. Guner, and J. Knowles (2002). Why do women wait? matching, wage
inequality, and the incentives for fertility delay. Review of Economic Dynamics 5,
815–855.
Chambers, M., C. Garriga, and D. E. Schlagenhauf (2005). Accounting for changes
in the homeownership rate. Unpublished manuscript.
Dynan, K., D. W. Elmendorf, and D. E. Sichel (2006). Can financial innovation
help to explain the reduced volatility of economic activity? Journal of Monetary
Economics Forthcoming.
Edelberg, W. (2003). Risk-based pricing of interest rates in household loan markets.
Unpublished manuscript.
Fisher, J. D. and S. Quayyum (2006). The great turn-of-the-century housing boom.
Federal Reserve Bank of Chicago Economic Perspectives Third Quarter, 22–37.
Floden, M. (2007). A note on the accuracy of markov-chain approximations to highly
persistent ar(1)-processes. Economics Letters forthcoming.
Florida, R. L. (Ed.) (1986). Housing and the New Financial Markets. Center for
Urban Policy Research.
Gerardi, K., H. S. Rosen, and P. Willen (2007). Do households benefit from financial
deregulation and innovation? the case of the mortgage market.
Gilbert, R. A. (1986). Requiem for regulation q: What it did and why it passed away.
Federal Reserve Bank of New York Review February, 22–37.
Gilbert, R. A. (1995). Inside the black box: The credit channel of monetary transmission. Journal of Economic Perspectives 9 (4), 27–48.
42

Haurin, D. R., P. H. Hendershott, and S. M. Wachter (1996). Wealth accumulation
and housing choices of young households: An exploratory investigation. Journal of
Housing Research 7 (1), 33–57.
Heathcote, J., K. Storesletten, and G. Violante (2004). The cross-sectional implications of rising wage inequality in the united states.
Jaffee, D. M., K. T. Rosen, B. M. Friedman, and L. R. Klein (1979). Mortgage
credit availability and residential construction. Brookings Papers on Economic
Activity 1979 (2), 333–386–37.
McCarthy, J. and R. W. Peach (2002). Monetary policy transmission to residential
investment. Federal Reserve Bank of New York Economic Policy Review May,
139–158.
Mitchell, O. S. and J. W. R. Phillips (2006). Social security replacement rates for alternative earnings benchmarks. University of Michigan Retirement Research Center
WP 2006-116.
Ryding, J. (1990). Housing finance and the transmission of monetary policy. Federal
Reserve Bank of New York Quarterly Review Summer, 42–55.
Stock, J. H. and M. W. Watson (2002). Has the business cycle changed and why?
Unpublished manuscript.
Storesletten, K., C. I. Telmer, and A. Yaron (2004). Cyclical dynamics in idiosyncratic
labor-market risk. Journal of Political Economy 112 (3), 695–717.
Tauchen, G. and R. Hussey (1991). Quadrature-based methods for obtaining approximate solutions to nonlinear asset pricing models. Econometrica 59 (2), 371–396.
Van-Order, R. (2000). The U.S. mortgage market: A model of dueling charters.
Journal of Housing Research 11 (2), 233–255.
Zagorsky, J. L. (1999). Young baby boomers’ wealth.
Wealth 45 (2), 135–156.

43

Review of Income and

Working Paper Series
A series of research studies on regional economic issues relating to the Seventh Federal
Reserve District, and on financial and economic topics.
Standing Facilities and Interbank Borrowing: Evidence from the Federal Reserve’s
New Discount Window
Craig Furfine

WP-04-01

Netting, Financial Contracts, and Banks: The Economic Implications
William J. Bergman, Robert R. Bliss, Christian A. Johnson and George G. Kaufman

WP-04-02

Real Effects of Bank Competition
Nicola Cetorelli

WP-04-03

Finance as a Barrier To Entry: Bank Competition and Industry Structure in
Local U.S. Markets?
Nicola Cetorelli and Philip E. Strahan

WP-04-04

The Dynamics of Work and Debt
Jeffrey R. Campbell and Zvi Hercowitz

WP-04-05

Fiscal Policy in the Aftermath of 9/11
Jonas Fisher and Martin Eichenbaum

WP-04-06

Merger Momentum and Investor Sentiment: The Stock Market Reaction
To Merger Announcements
Richard J. Rosen

WP-04-07

Earnings Inequality and the Business Cycle
Gadi Barlevy and Daniel Tsiddon

WP-04-08

Platform Competition in Two-Sided Markets: The Case of Payment Networks
Sujit Chakravorti and Roberto Roson

WP-04-09

Nominal Debt as a Burden on Monetary Policy
Javier Díaz-Giménez, Giorgia Giovannetti, Ramon Marimon, and Pedro Teles

WP-04-10

On the Timing of Innovation in Stochastic Schumpeterian Growth Models
Gadi Barlevy

WP-04-11

Policy Externalities: How US Antidumping Affects Japanese Exports to the EU
Chad P. Bown and Meredith A. Crowley

WP-04-12

Sibling Similarities, Differences and Economic Inequality
Bhashkar Mazumder

WP-04-13

Determinants of Business Cycle Comovement: A Robust Analysis
Marianne Baxter and Michael A. Kouparitsas

WP-04-14

The Occupational Assimilation of Hispanics in the U.S.: Evidence from Panel Data
Maude Toussaint-Comeau

WP-04-15

1

Working Paper Series (continued)
Reading, Writing, and Raisinets1: Are School Finances Contributing to Children’s Obesity?
Patricia M. Anderson and Kristin F. Butcher

WP-04-16

Learning by Observing: Information Spillovers in the Execution and Valuation
of Commercial Bank M&As
Gayle DeLong and Robert DeYoung

WP-04-17

Prospects for Immigrant-Native Wealth Assimilation:
Evidence from Financial Market Participation
Una Okonkwo Osili and Anna Paulson

WP-04-18

Individuals and Institutions: Evidence from International Migrants in the U.S.
Una Okonkwo Osili and Anna Paulson

WP-04-19

Are Technology Improvements Contractionary?
Susanto Basu, John Fernald and Miles Kimball

WP-04-20

The Minimum Wage, Restaurant Prices and Labor Market Structure
Daniel Aaronson, Eric French and James MacDonald

WP-04-21

Betcha can’t acquire just one: merger programs and compensation
Richard J. Rosen

WP-04-22

Not Working: Demographic Changes, Policy Changes,
and the Distribution of Weeks (Not) Worked
Lisa Barrow and Kristin F. Butcher

WP-04-23

The Role of Collateralized Household Debt in Macroeconomic Stabilization
Jeffrey R. Campbell and Zvi Hercowitz

WP-04-24

Advertising and Pricing at Multiple-Output Firms: Evidence from U.S. Thrift Institutions
Robert DeYoung and Evren Örs

WP-04-25

Monetary Policy with State Contingent Interest Rates
Bernardino Adão, Isabel Correia and Pedro Teles

WP-04-26

Comparing location decisions of domestic and foreign auto supplier plants
Thomas Klier, Paul Ma and Daniel P. McMillen

WP-04-27

China’s export growth and US trade policy
Chad P. Bown and Meredith A. Crowley

WP-04-28

Where do manufacturing firms locate their Headquarters?
J. Vernon Henderson and Yukako Ono

WP-04-29

Monetary Policy with Single Instrument Feedback Rules
Bernardino Adão, Isabel Correia and Pedro Teles

WP-04-30

2

Working Paper Series (continued)
Firm-Specific Capital, Nominal Rigidities and the Business Cycle
David Altig, Lawrence J. Christiano, Martin Eichenbaum and Jesper Linde

WP-05-01

Do Returns to Schooling Differ by Race and Ethnicity?
Lisa Barrow and Cecilia Elena Rouse

WP-05-02

Derivatives and Systemic Risk: Netting, Collateral, and Closeout
Robert R. Bliss and George G. Kaufman

WP-05-03

Risk Overhang and Loan Portfolio Decisions
Robert DeYoung, Anne Gron and Andrew Winton

WP-05-04

Characterizations in a random record model with a non-identically distributed initial record
Gadi Barlevy and H. N. Nagaraja

WP-05-05

Price discovery in a market under stress: the U.S. Treasury market in fall 1998
Craig H. Furfine and Eli M. Remolona

WP-05-06

Politics and Efficiency of Separating Capital and Ordinary Government Budgets
Marco Bassetto with Thomas J. Sargent

WP-05-07

Rigid Prices: Evidence from U.S. Scanner Data
Jeffrey R. Campbell and Benjamin Eden

WP-05-08

Entrepreneurship, Frictions, and Wealth
Marco Cagetti and Mariacristina De Nardi

WP-05-09

Wealth inequality: data and models
Marco Cagetti and Mariacristina De Nardi

WP-05-10

What Determines Bilateral Trade Flows?
Marianne Baxter and Michael A. Kouparitsas

WP-05-11

Intergenerational Economic Mobility in the U.S., 1940 to 2000
Daniel Aaronson and Bhashkar Mazumder

WP-05-12

Differential Mortality, Uncertain Medical Expenses, and the Saving of Elderly Singles
Mariacristina De Nardi, Eric French, and John Bailey Jones

WP-05-13

Fixed Term Employment Contracts in an Equilibrium Search Model
Fernando Alvarez and Marcelo Veracierto

WP-05-14

Causality, Causality, Causality: The View of Education Inputs and Outputs from Economics
Lisa Barrow and Cecilia Elena Rouse

WP-05-15

3

Working Paper Series (continued)
Competition in Large Markets
Jeffrey R. Campbell

WP-05-16

Why Do Firms Go Public? Evidence from the Banking Industry
Richard J. Rosen, Scott B. Smart and Chad J. Zutter

WP-05-17

Clustering of Auto Supplier Plants in the U.S.: GMM Spatial Logit for Large Samples
Thomas Klier and Daniel P. McMillen

WP-05-18

Why are Immigrants’ Incarceration Rates So Low?
Evidence on Selective Immigration, Deterrence, and Deportation
Kristin F. Butcher and Anne Morrison Piehl

WP-05-19

Constructing the Chicago Fed Income Based Economic Index – Consumer Price Index:
Inflation Experiences by Demographic Group: 1983-2005
Leslie McGranahan and Anna Paulson

WP-05-20

Universal Access, Cost Recovery, and Payment Services
Sujit Chakravorti, Jeffery W. Gunther, and Robert R. Moore

WP-05-21

Supplier Switching and Outsourcing
Yukako Ono and Victor Stango

WP-05-22

Do Enclaves Matter in Immigrants’ Self-Employment Decision?
Maude Toussaint-Comeau

WP-05-23

The Changing Pattern of Wage Growth for Low Skilled Workers
Eric French, Bhashkar Mazumder and Christopher Taber

WP-05-24

U.S. Corporate and Bank Insolvency Regimes: An Economic Comparison and Evaluation
Robert R. Bliss and George G. Kaufman

WP-06-01

Redistribution, Taxes, and the Median Voter
Marco Bassetto and Jess Benhabib

WP-06-02

Identification of Search Models with Initial Condition Problems
Gadi Barlevy and H. N. Nagaraja

WP-06-03

Tax Riots
Marco Bassetto and Christopher Phelan

WP-06-04

The Tradeoff between Mortgage Prepayments and Tax-Deferred Retirement Savings
Gene Amromin, Jennifer Huang,and Clemens Sialm

WP-06-05

Why are safeguards needed in a trade agreement?
Meredith A. Crowley

WP-06-06

4

Working Paper Series (continued)
Taxation, Entrepreneurship, and Wealth
Marco Cagetti and Mariacristina De Nardi

WP-06-07

A New Social Compact: How University Engagement Can Fuel Innovation
Laura Melle, Larry Isaak, and Richard Mattoon

WP-06-08

Mergers and Risk
Craig H. Furfine and Richard J. Rosen

WP-06-09

Two Flaws in Business Cycle Accounting
Lawrence J. Christiano and Joshua M. Davis

WP-06-10

Do Consumers Choose the Right Credit Contracts?
Sumit Agarwal, Souphala Chomsisengphet, Chunlin Liu, and Nicholas S. Souleles

WP-06-11

Chronicles of a Deflation Unforetold
François R. Velde

WP-06-12

Female Offenders Use of Social Welfare Programs Before and After Jail and Prison:
Does Prison Cause Welfare Dependency?
Kristin F. Butcher and Robert J. LaLonde
Eat or Be Eaten: A Theory of Mergers and Firm Size
Gary Gorton, Matthias Kahl, and Richard Rosen
Do Bonds Span Volatility Risk in the U.S. Treasury Market?
A Specification Test for Affine Term Structure Models
Torben G. Andersen and Luca Benzoni

WP-06-13

WP-06-14

WP-06-15

Transforming Payment Choices by Doubling Fees on the Illinois Tollway
Gene Amromin, Carrie Jankowski, and Richard D. Porter

WP-06-16

How Did the 2003 Dividend Tax Cut Affect Stock Prices?
Gene Amromin, Paul Harrison, and Steven Sharpe

WP-06-17

Will Writing and Bequest Motives: Early 20th Century Irish Evidence
Leslie McGranahan

WP-06-18

How Professional Forecasters View Shocks to GDP
Spencer D. Krane

WP-06-19

Evolving Agglomeration in the U.S. auto supplier industry
Thomas Klier and Daniel P. McMillen

WP-06-20

Mortality, Mass-Layoffs, and Career Outcomes: An Analysis using Administrative Data
Daniel Sullivan and Till von Wachter

WP-06-21

5

Working Paper Series (continued)
The Agreement on Subsidies and Countervailing Measures:
Tying One’s Hand through the WTO.
Meredith A. Crowley

WP-06-22

How Did Schooling Laws Improve Long-Term Health and Lower Mortality?
Bhashkar Mazumder

WP-06-23

Manufacturing Plants’ Use of Temporary Workers: An Analysis Using Census Micro Data
Yukako Ono and Daniel Sullivan

WP-06-24

What Can We Learn about Financial Access from U.S. Immigrants?
Una Okonkwo Osili and Anna Paulson

WP-06-25

Bank Imputed Interest Rates: Unbiased Estimates of Offered Rates?
Evren Ors and Tara Rice

WP-06-26

Welfare Implications of the Transition to High Household Debt
Jeffrey R. Campbell and Zvi Hercowitz

WP-06-27

Last-In First-Out Oligopoly Dynamics
Jaap H. Abbring and Jeffrey R. Campbell

WP-06-28

Oligopoly Dynamics with Barriers to Entry
Jaap H. Abbring and Jeffrey R. Campbell

WP-06-29

Risk Taking and the Quality of Informal Insurance: Gambling and Remittances in Thailand
Douglas L. Miller and Anna L. Paulson

WP-07-01

Fast Micro and Slow Macro: Can Aggregation Explain the Persistence of Inflation?
Filippo Altissimo, Benoît Mojon, and Paolo Zaffaroni

WP-07-02

Assessing a Decade of Interstate Bank Branching
Christian Johnson and Tara Rice

WP-07-03

Debit Card and Cash Usage: A Cross-Country Analysis
Gene Amromin and Sujit Chakravorti

WP-07-04

The Age of Reason: Financial Decisions Over the Lifecycle
Sumit Agarwal, John C. Driscoll, Xavier Gabaix, and David Laibson

WP-07-05

Information Acquisition in Financial Markets: a Correction
Gadi Barlevy and Pietro Veronesi

WP-07-06

Monetary Policy, Output Composition and the Great Moderation
Benoît Mojon

WP-07-07

Estate Taxation, Entrepreneurship, and Wealth
Marco Cagetti and Mariacristina De Nardi

WP-07-08

6

Working Paper Series (continued)
Conflict of Interest and Certification in the U.S. IPO Market
Luca Benzoni and Carola Schenone
The Reaction of Consumer Spending and Debt to Tax Rebates –
Evidence from Consumer Credit Data
Sumit Agarwal, Chunlin Liu, and Nicholas S. Souleles

WP-07-09

WP-07-10

Portfolio Choice over the Life-Cycle when the Stock and Labor Markets are Cointegrated
Luca Benzoni, Pierre Collin-Dufresne, and Robert S. Goldstein

WP-07-11

Nonparametric Analysis of Intergenerational Income Mobility
with Application to the United States
Debopam Bhattacharya and Bhashkar Mazumder

WP-07-12

How the Credit Channel Works: Differentiating the Bank Lending Channel
and the Balance Sheet Channel
Lamont K. Black and Richard J. Rosen

WP-07-13

Labor Market Transitions and Self-Employment
Ellen R. Rissman

WP-07-14

First-Time Home Buyers and Residential Investment Volatility
Jonas D.M. Fisher and Martin Gervais

WP-07-15

7