Jack B Homer.

Home insurance in a changing residential community : a system dynamics approach and case study online

. (page 1 of 5)
Online LibraryJack B HomerHome insurance in a changing residential community : a system dynamics approach and case study → online text (page 1 of 5)
Font size
QR-code for this ebook

c. i)-







Jack B. Homer

WP 110$-80

December 1979








Jack B. Homer

WP llOg-80 December 1979

System Dynamics Group
Alfred P. Sloan School of Management
Massachusetts Institute of Technology
Cambridge, Massachusetts





1.1 The Problem 1

1 . 2 The Approach 3



2.1 An Introduction to the Model 7

2.2 The Population Sector 11

2.3 The Housing Sector 14

2.4 The Quality of Neighborhood Life Sector 17

2.5 The Insurance Sector 21


3.1 History of Melrose 25

3.2 Some Facts Concerning Melrose 28

3.3 Parameters Used to Describe Melrose 29


4. 1 Base Run Description 37

4.2 A Theory of Community Change 44

4.3 The Self-Sustaining Transition 52

4.4 The Role of Insurance Availability 57





D-3 1 80



This report examines transition in inner suburban communities and
how insurance companies should regard the sometimes rapid changes that
occur in these areas. In order to understand how the physical, economic,
and social aspects of a community work together to create a self-sustaining
transition that produces losses for the insurer of homes, a system dynamics
computer simulation model of home insurance coverage in a residential
community was constructed.

The model, called INSUR2, contains four major sectors: popula-
tion, housing, quality of neighborhood life, and insurance. There are
three levels of population — upper, middle, and lower class. These levels
are affected by rates of birth and death and in-migration and out-migration
which are generated within the model. There are six types of housing which
differ according to market values and replacement costs. Housing construc-
tion, demolition, obsolescence, and renovation are endogenously generated
and affect these housing stocks. The quality of neighborhood life is a
central concept to this study and differentiates it from other models which
consider almost exclusively the physical or economic bases of change.
Quality of neighborhood life reflects community services as well as the
social stability of the local neighborhood. The insurance sector keeps
track of policies which are written and then may be cancelled or
non-renewed. Mainly on the basis of neighborhood conditions and past
performance in the area, the company decides to what extent it will make
insurance available in the community. Profits in the community are com-



puted from preniums earned minus expenses. Insurable property damages
include property crimes and some wear and tear. Property crimes are
divided into the two categories of arson and "miscellaneous" property
crimes. Housing conditions in the simulated community are determined by
looking at the accumulation of unrepaired damages.

The model was parameterized to represent the city of Melrose,
Massachusetts, which is located about seven miles north of Boston and is a
typical inner suburb which has undergone considerable changes in population
and housing since the 19403. Simulation of the Melrose-adapted model
produces a history starting in 1900 which closely matches available data
and descriptive material. It portrays a future for Melrose through the
year 2020 in which the decline of the last forty years may become sustained
and therefore cause some worries among insurers.

The INSUR2 model was analyzed in terms of its feedback structure
to discover why the decline occurs and what can be done about it. The two
major results concerning the mechanism of decline are:

(1 ) The transition is triggered by decreasing availability of choice

parcels of land, which leads to decreased housing construction and
increased conversions to multi-family, renter-occupied, lower-
class dwellings;
(2) In the case of an originally well-to-do community like Melrose,

the downward transition — which implies a decline in company profit



margin — is only sustained (into ultimate decay and abandonment) if
the changing population mix leads to inter-class tensions and
defensiveness which break down neighborhood cohesiveness.
Neighborhood cohesiveness is thus identified as the key indicator of a
community's health and its ability to maintain the stability required to
support good home maintenance and protection against various damages.
Neighborhood cohesiveness is defined as the presence of good neighborly
relations and the common feeling that the neighborhood is relatively free
of divisiveness .

A nunber of simulations consider alternative futures for Melrose
and similar communities. The results are:

o The "return-to-the-city movement" which has led to gentrification
in some inner cities will not benefit Melrose significantly and
may even prove marginally detrimental to company profits in the
short run.
o A complete cutoff of insurance availability accelerates the
transition by reducing the fraction of homes that are owner-
occupied .
If, instead of the partial insurance availability withdrawal of

the base run (about 507o by the year 2020), insurance is made fully
available, the effect on the community is negligible and the
effect on profits is somewhat negative after 1990.
o The housing programs which are most effective in slowing down and
eventually reversing neighborhood decline still produce little
noticeable improvement over the base run until 1990, after which



they push the community back up to its past (I960) mix of
population and housing primarily by displacing lower-class
residents from the community.

o Programs aimed at improving neighborhood services and facilities
can improve community satisfaction and pride, but may not have
lasting effects, because they do not directly improve social
cohesiveness. However, they are worth advocating as part of a
comprehensive revitalization program, and they directly increase
company profits by increasing local commitment. The positive
image-building value of historic preservation is noted.

o The promotion of strong community organizations is identified as
an effective way of stabilizing the neighborhoods and increasing
the quality of life (and therefore, company profits). Short-term
investments in developing capable organizations can pay off in the
long term, because such groups attract citizen participation and
help create a more cohesive and committed community.

In the Conclusions, it is recognized that an insurance company is
not generally in the position to coordinate or implement the beneficial
programs mentioned above. As one participant in the community, however, it
can influence what happens and persuade the city government and other
important participants to take the kinds of actions which will make company
withdrawal unnecessary.

D-3 1 80

To the memory of Gil Low, a dear friend, advisor, and teacher

D-3180 1


1. 1 The Problem

There are many communities in the United States that have
experienced profound changes during the last few decades. Metropolitan
residential patterns are now changing more quickly than ever, for a variety
of reasons, among them the steady saturation of both urban and suburban
land areas, the increasingly significant needs and preferences of the baby
boom generation, the high and rising costs of new housing and private
transportation, and the generally fluid nature of American society.

Insurance companies, as well as banks and other financial
institutions, have a special interest in the destinies of communities.
They provide important services for their customers and hope to make a
profit in return. In this way, communities are considered markets for
investment by such institutions. When the risk and uncertainty associated
with an investment is high, the potential investor may have second thoughts
about entering the market and the present investor may wish to pull out of
it. The deterioration of many inner city areas and the resulting losses to
some insurance companies and banks have sensitized both industries to the
dangers associated with rapidly changing residential areas.

A number of suburban areas, especially those located closest to
the city, have been in the midst of the changes in this country mentioned
above. An investor's natural reaction is to regard these changes with
suspicion. Some observers believe the inner suburbs will follow the same
path of deterioration as the inner cities. This statement often reflects

D-3180 2

the notion that all so-called "transitional" neighborhoods go through a
fixed sequence of five or six well-defined stages of deterioration that

begins with a decline in socioeconomic status and ends with decay and

abandonment. Since many of the inner suburbs have aging housing stocks

and are being increasingly "invaded" by residents of lower social status,
insurance companies and banks may be tempted to get out before such
communities "tip" and become unprofitable.

This paper will focus on the involvement of insurance companies in
transitional inner suburban areas. The first consideration will be whether
the hesitancy about these areas is at all justified. After all, "neighbor-
hoods in fact change all the time, and while they might decline, they

also possess the power to regenerate themselves." It is likely that the

worries are justified in some cases and not in others. Therefore, insur-
ance companies should have an appreciation for the leading indicators or
preconditions which point to an extended period of losses with no ultimate
recovery of profits.

The second consideration will be what an insurance company should
do if it determines that a given neighborhood is a bad risk. If the
canpany does not presently insure homes in the area, the answer is probably
to stay out; that is the company's prerogative, as long as it can be
demonstrated that the decision to stay out is not based on the kinds of
discrimination (such as racial discrimination) that have been shown to
violate the Fair Housing Act. However, if the company is already active
in the community, its range of realistic alternatives may be narrowed by
public interest considerations.

D-3180 3

When an insurance company agrees to insure homes in a community,
it enters into a tacit agreement to act with the community's welfare in
mind. In cases of policyholder fraud or arson or other breach of contract,
it is in both the company's interest and in the community's greater
interest to cancel individual insurance policies. But to simply withdraw
from a community because of the fear of future losses is socially
irresponsible if the withdrawal condeiTins the community to severe
deterioration it need not otherwise have faced. The company should
therefore make an effort to work together with various public and private
agencies and individuals both inside and outside the community to make a
disruptive and damaging exit unnecessary. Aside from the obvious virtues
of such actions, the long-run results might well include a financial return
to the insurance company that makes the effort worthwhile. This study will
explore the relative merits of the various actions that might be taken in
an attempt to stabilize or revitalize a transitional community.

1.2 The Approach

The uncertainties of insurance companies concerning transitional
inner suburbs and how they should be treated cannot be transformed into
certainties by any methodology or analytic approach. The purpose of this
paper is rather to enhance understanding of the problem and pave the way
for informed decision-making. Even though different communities call for
different types and degrees of institutional response, the major tenet here
is that these differences can all be understood within one theoretical
framework. The basis for this framework should be the essential elements

D-3 1 80 4

found in every community, such as a population, a housing stock, and
community services. The important decisions and forces that affect these
elements and that are affected by them should also be represented; for
example, the decision to move out of the community or the decision to
renovate a house. By examining a simplified but still realistic model of a
typical canmunity, one hopes to draw certain conclusions about why the
regularities and similarities present in a wide variety of settings exist

and which policy levers may be the most useful in generating beneficial


This paper discusses a generic system dynamics model of a
residential community and its interactions with an insurance company (or

group of insurance companies) and with the rest of its "limitless

environment". The premise in using such a model is that a detailed

analysis of the decision-making structure of a system is more fruitful in
understanding its behavior than any analogical model, such as the
biological model of urban decline which "assumes that neighborhoods go
through an irreversible cycle of growth, maturity, decline, and death which
varies only in duration and intensity". Such analogical models neither
demonstrate why the observed behavior follows from the underlying structure
nor do they suggest how policymakers should act.

System dynamics emphasizes the importance of feedback in all
decision-making processes. For exanple, building contractors will
generally respond to excess housing demand as reflected in market values by
building new dwellings. Other things staying the same, construction of
homes will decrease the excess demand for them and, in turn, decrease
housing construction. The situation can be illustrated in the "causal-
loop" format shown in Figure 1. Each arrow in the loop shows the direction

D-3 1 80


Housing Demand



Figure 1: A feedback loop involving housing construction,
excess demand, and price.

of causation and the plus or minus sign at the arrowhead indicates the
direction of effect for a given causal link. Thus, the positive link
between excess housing deinand and housing price says that if the former
increases, so will the latter (other things still being equal). The loop
symbol in the middle of the diagram indicates the overall polarity of the
loop, in this case negative. A negative loop can be thought of as a
controlling or correcting loop, in that pressures (such as excess housing
demand) elicit a response within the loop (housing construction) which
tends to diminish those pressures. A positive loop, on the other hand, can
be responsible for runaway growth or collapse; a standard example of
positive loop behavior is exponential population growth, which can be

traced back to the fact that more people means more babies means more

people and so on .

The computer simulation model used in this study is called INSUR2.

Such a model is necessary to draw the correct behavioral conclusions from

an assumed set of structural relationships as complex as those being

considered here. INSUR2 can be parameterized to represent a wide diversity

of communities and insurance companies. To study the specific problem of

D-3180 6

inner suburban decline, a case study approach has been taken. The city of
Melrose, Massachusetts, was chosen as a community which is typical of the
kind of suburban areas under consideration. The conclusions that are
reached relative to Melrose should apply to other transitional residential
canmunities as well.

Various kinds of available information on Melrose and its
surroundings and on the property insurance business were used to develop
and parameterize the model and help establish the reasonableness of its
output. This data came from a variety of sources, both published and
personally communicated. The Corporate Research and Personal Lines
divisions of Commercial Union and the Planning Office of Melrose were
especially helpful in making this project possible.


2.1 An Introduction to the Model
Scope and Boundary

The INSUR2 model can be used to examine the internal dynamics of a
residential community and how they affect and are affected by insurance
canpanies that offer homeowner's coverage.*

A residential or "bedroom" community has relatively few commercial
or industrial structures, and most of its working residents are employed
outside the community. The community has a fixed geographical boundary
beyond which it cannot expand. People arrive to and depart from the
community across this boundary; and although their internal movements are
not computed in the model, the broader concept of residential stability is
made explicit.

The population in INSUR2 is subdivided according to social class
but not age, sex, religion, or race. The housing stock is subdivided
according to market value and replacement cost and is subject to damage and
repair. The community contains an unspecified number of sub-neighborhoods
which have caninon access to community services. Services and stability are
the two factors in the model which directly determine the community's
quality of life.

* System Dynamics Group Memorandum D-3180 includes a Technical Appendix
with canplete model documentation.

D-3 1 80 8

In reality, a community is often affected by events determined
outside of its borders, including broad economic and demographic changes
and more localized events, like the construction of new transportation
lanes. These events are considered "exogenous" or beyond the scope of the
model, but they can be assessed in a general way by looking at their
primary effects on population and housing flows. (In this way, the
"return-to-the-city movement" is examined in Section 5.)

The insurance company has a limited but realistic role in the
model. It can determine homeowner's insurance availability in a given
community; in real life, this is done by hiring, firing, or reallocating
insurance agents and by setting standards that control the type of business
that is accepted. The setting of premiums is beyond the scope of the
model, however, since insurance rates are not determined on a
community-by-community basis.


Figure 2 provides a pictorial overview of the INSUR2 model. The
community develops within a fixed land area and consists of three
interconnected sectors: population, housing, and quality of neighborhood
life. The community is linked to a fourth sector, the insurance sector.

The land area is used in computing the population density and the
fraction of land still available for residential development. In addition,
zoning restrictions arise from the need to parcel out the limited land for
various uses. The population density affects the quality of neighborhood
life, because the distance between people and the proximity of group


I Insurance

I Sector





Land Area

Figure 2: An Overview of INSUR2

D-3180 10


territories can affect neighborliness and crime. Land availability and
zoning can have critical effects on housing activities. As the area
becomes fully developed, choice parcels of land become relatively scarce
and the cost of land rises relative to property values; under such
conditions, new construction is inhibited and demolition and renovation are


encouraged .

The following is a brief description of the numbered linkages in
Figure 2:

(1 ) The availability and price of housing affect population in-migration
and out-migration;

(2) The size and composition of the population affect housing
construction, demolition, obsolescence, and renovation;

(3) The quality of neighborhood life (an aggregate measure of the
vitality, robustness, or strength of the community's sub-
neighborhoods) is affected by population mix, population density, and

(4) Quality of neighborhood life is also affected by the adequacy of the
housing stock, its condition, and the fraction of homes which are

(5) The quality of neighborhood life is an important factor in the
determination of both property damage (including property crime) and
repair as well as new housing construction;

(6) Quality of neighborhood life also affects people's decisions to enter
or leave the community;

D-3180 11

(7) Insurance companies are affected both directly and indirectly by all
three community sectors. Property danage affects profits in an
obvious way. When making decisions that affect the writing of new
policies and the non-renewal or cancellation of existing policies,
insurance companies look not only to past experience with the
community, but also to such indicators of future risk as the condition
and cost of housing and the quality of neighborhood life;

(8) The availability of property insurance affects the entire community.
People who desire insurance coverage to limit their liability or need
insurance to obtain a mortgage are deterred from entering if they
cannot get it. Arson-for-profit is possible only when a building is
insured; and when it is perceived that cancellations and non-renewals
are on the increase, property speculators are often tempted to burn
their buildings quickly and get out before the local housing market
slumps. Finally, the lack of home insurance will tend to magnify the
residents' feelings of insecurity or fear due to property crime, since
they will have to pay for needed repairs out of their own pockets.

2.2 The Population Sector

Figure 3 shows the basic structure of the population sector, using
conventional system dynamics symbols to represent levels and rates.* There
are three levels of population: upper class, middle class, and lower

* A level ( I I ) represents an accunulation or integration of rates
( "y^ * ) . In mathematical notation:





Net Blr



.' .)


Upper class






Net Birth^






^ i
Net Births

M >

Middle class



Lower class
I Population

Departures i

Figure 3: Population Sector

D-3180 13

class. Each population level is changed by a net birth rate, an arrival
(in-niigration) rate, and a departure (out-migration) rate. It is assuned
for the sake of simplicity that present inhabitants cannot move from one
class to another.

The model identifies these separate classes of people for several
reasons. First, well-to-do people can afford to purchase and maintain
premium single- family dwellings, whereas lower class people tend to be
renters and live in old or inexpensive multi-family dwellings. Second,
income is closely related to the quality, quantity, and specific kinds of
services provided in the community. Third, social class is connected with
the kind of neighboring that is observed: "Greater economic well-being
decreases the need for mutual aid and increases the use of critical,
selective faculties ... In the words of one writer, 'the higher the level of
prosperity, the higher the fences.'" Fourth, different socioeconomic
classes are generally associated with different levels of education, types
of families, and sets or ranges of skills, material possessions, interests,
and values. These attributes are important to the extent that they create
relatively self-contained subgroups within the community which may view
each other with suspicion and uncertainty. This is most noticeable when
the different classes correspond to different races or ethnicities, which
is canmonly the case in American cities.

Within the population sector, the social mix is an important
factor in the decision to move into or out of a community. More broadly,
"[The characteristics of other residents] may determine how people will
react to the adequacy of their houses and facilities, whether they intend

to stay or move away, and how they cope with noise, overcrowding, and other

1 1
mconv eniences ."

D-3180 14

2.3 The Housing Sector

Figure 4 shows the basic configuration of the housing sector.

1 3 4 5

Online LibraryJack B HomerHome insurance in a changing residential community : a system dynamics approach and case study → online text (page 1 of 5)