Get documents about "Efficiency
Document Sample
EFFICIENCY OF THE GERMAN AND OF THE AUSTRIAN INSURANCE INDUSTRY – A
COMPARISON
Randy I. Anderson*
Karl C. Ennsfellner**
Jochen Russ***
*
Dr. Randy I. Anderson is a Vice President in the Investment Research Department of Prudential Real
Estate Investors. The Prudential Insurance Company of America, 751 Broad Street, Newark,
NJ 07102-3777, Phone: 001 973 734 1308, e-mail: Randy.Anderson@prudential.com
**
Dr. Karl C. Ennsfellner is Assistant Professor of Insurance at the Department of Risk Management
and Insurance at the Vienna University of Economics and Business Administration, Augasse 2 – 6,
A-1090 Vienna, Austria. Phone: +43 1 31336-4692, e-mail: Karl.Ennsfellner@wu-wien.ac.at
***
Dr. Jochen Russ is Associate Director of the Institut fuer Finanz- und Aktuarwissenschaften and
lecturer at the University of Ulm, Helmholtzstr. 22, 89081 Ulm, Germany.
Phone: +49 731 50 31234, e-mail: j.russ@ifa-ulm.de
ABSTRACT
In this study we examine the efficiency of the insurance markets of Germany and Austria. We estimate
input and output efficiency for the year 1998 using Data Envelopment Analysis (DEA). In addition we
compute the contribution of the inputs and outputs to the total inefficiencies of the firms not operation
on the efficient frontier. We found significant differences between the countries regarding output
efficiencies in the non-life insurance market as well as regarding output and input efficiencies of the
life insurance business. Furthermore our results suggest that there are also differences in the inputs and
outputs that contribute to the inefficiencies on the German and on the Austrian insurance market.
I. INTRODUCTION
The European single insurance market has almost completed the harmonization of its legal framework.
By introducing the Euro, the European Union removed another important barrier to cross border trade
between the member states. However, experts believe that differences in the language are one of the
most important obstacles for cross-border activities in the personal lines business on the European
single insurance market. Sharing the same language, this barrier does not exist between Germany and
Austria.
Despite the differences in size, a number of additional similarities can be observed on the two markets.
For more than a century, the approach to the regulatory law of the German and the Austrian insurance
market has been similar leading to a comparable insurance culture. The social insurance system with
the dominance of the state-provided old-age pension system or the strong state-provided health care
insurance system should be mentioned as another example for the similarities in the environment of the
insurance business. Furthermore, the insurance technology of these countries is about comparable.
Following the implementation of the common currency1, it can be expected that the similarities of the
markets lead to an increase in cross-border activities between Germany and Austria. In addition,
following the deregulation of the German and Austrian insurance industries, domestic competition
increased. Taking these facts into account, it is crucial for insurance companies to produce as
efficiently as possible to survive in such an environment. Differences in efficiency of insurance
companies in the member states of European single insurance market largely will determine the
structure of the national insurance industries. This is especially true for Austria, a small market
compared to Germany, being among the three largest insurance markets of the European single
insurance market.
With this background our paper concentrates on the comparison of the efficiency of insurance
companies in Germany and in Austria. Our paper is organized in six sections. After the introduction,
we present a brief overview of the existing literature to the topic (section II). Due to the lack of
international efficiency comparisons of insurance markets, we extend our review to the relevant
banking literature. Section III describes briefly the insurance markets of Germany and of Austria.
Section IV deals with the data and the methodology used in our study. The results are presented in
Section V, while section VI summarizes our findings and gives directions for future research.
II. LITERATURE REVIEW
Our literature research has revealed a lack of studies that exclusively concentrate on the comparison of
the efficiency of the insurance industry of different countries. However, we found one study by Fecher
and Pestieu (1993) that included the insurance industry in a broader picture. In their paper, Fecher and
Pestieu present estimates of multifactor productivity change, decomposed in technical progress and
efficiency change, for financial services markets. Using a fixed effects model based on a Distribution
Free Approach (DFA), the authors analyze the development of the efficiency of the financial services
markets in 11 OECD countries between 1971 and 1986. The study also looks at what extent efficiency
changes can be explained in part by the strength of regulation and by the competitive conditions in the
various countries.
1
At present, the Euro is used as book money in eleven member states of the European Union. Notes
and coins will circulate starting on January 1, 2002. It is planned to complete the transformation of the
currencies by February 28, 2002.
1
Extending the literature research to banking studies we found six studies in this field. The most recent
study by Dietsch and Lozano-Vivas (2000) compares the efficiency of the Spanish and the French
banking industry using DFA for data between 1988 and 1992. In this study, the authors integrate
environmental variables into the definition of the common frontier. The result of the study shows that
the Spanish banks seem to suffer from structural disadvantages when compared to the French banking
industry. The findings reveal that the reasons for the excessive costs of the Spanish banks can be
identified in their particular environmental and regulatory conditions as compared to the French banks.
Allen and Rai (1996) present a study of banking systems of 15 countries, using DFA and the Stochastic
Frontier Approach (SFA). The authors use data for the period between 1988 to 1992 to compare X-
efficiency measures across countries distinguished by different regulatory environments. According to
the regulatory environment, countries are divided into countries that prohibit and in countries that allow
functional integration of commercial banking and investment banking. The findings suggest that large
banks in countries that prohibit functional integration of commercial banking and investment banking
have significantly less efficient operations than any other group of banks for the period under
investigation.
Pastor, Pérez and Quesada (1997) compare the efficiency of seven different European banking systems
and the US-banking system, using Data Envelopment Analysis (DEA). Like Dietsch and Lozano-
Vivas (2000), and like Fecher and Pestieu (1993), Pastor, Pérez and Quesada (1997) pool the cross-
country data to define a common frontier. Their result suggests that France, Spain and Belgium seem
to be the countries with the most efficient banking systems, whereas Austria, Germany and the United
Kingdom show the lowest efficiency levels. The problem associated with the definition of a common
frontier is that cross-country differences are not specified. This may affect the cross country results.
Berg, Forsund, Hjalmarsson and Suominen (1993) follow a different approach. Using DEA, they
compare the efficiency of the banking systems of Finland, Norway and Sweden in two steps. In a first
step the authors define separate frontiers for each country and performed pair-wise comparisons of the
countries using separate frontiers. In a second step they pooled the data to define a common frontier
and compare the results across the countries again. The results show Sweden to have the most efficient
banking system. A follow up-study by Berg, Bukh and Forsund (1995) with Denmark added to the
sample shows similar results. However, these studies suggest a different level of efficiency for Sweden
than found by Fecher and Pestieau (1993). In another study, Bergendahl (1995) covers the efficiency
of the banking system of Finland, Norway, Sweden and Denmark. In contrast to Berg, Bukh and
Forsund (1995) and to Berg, Forsund, Hjalmarsson and Suominen (1993), Bergendahl develops a
“reference bank” that is composed of the most efficient components of the banks of the countries
covered. Using the Mixed Optimal Strategy (MOS), leading to higher benchmarks than DEA,
Bergendahl shows results that might be possible for banks, but have not been achieved yet.
III. THE INSURANCE MARKETS
The German Insurance Market
Germany has a population of approximately 82 million. Prior to the German unification that took place
in 1990, about 65 million people lived in what then was called the Federal Republic of Germany or
West Germany and about 16 million in the German Democratic Republic or East Germany.
The Federal Supervisory Office for Insurance (BAV) is part of the Ministry of Finance. All private
insurance companies headquartered in Germany or on the market with branch offices in Germany are
subject to the BAV’s control unless they are registered in another member state of the European single
insurance market. After the deregulation of the German insurance market that took place in 1994, the
major assignment of the BAV is solvency control. According to the German Regulatory Insurance
Law, life and health insurance must be transacted by separate companies from those writing non-life.
However, it is allowed to perform these transactions in the same group.
Domestic insurers are allowed to conduct business only in the legal form of a stock company, of a
mutual company, or in the legal form of a public corporation. The latter are government agencies
subject to public law. They are owned by the regions or federal states in which they operate, either
directly or through the state savings banks and are regulated by the federal states. In 1998, there were
719 insurance companies regulated by the BAV and another 12 by the federal states. In Germany there
2
are 136 life insurance companies, 146 pension funds, 56 funeral expense companies, 60 health insurers,
283 non-life insurers, and 50 reinsurers. In 1997, foreign insurers from within the European single
insurance market had a market share of 0.6% of the premiums, 0.4% are apportioned to subsidiaries of
foreign companies, 0.2% to business written under the freedom to provide services act. The sum of
gross premiums earned by insurance companies in 1998 amounted to 127,420 million Euro,
representing about 6.5 % of Germany’s total GDP, where life insurance accounts for 41 %, health
insurance for 15.2% and non-life insurance (other than health) had a share of 41.8 % of the German
insurance market.
Concentration was the highest in health insurance, where the largest company accounted for 14 % of
the premiums, the largest three companies controlled 39.9 % of the health insurance market and the
largest 10 companies controlled 70 % of the premiums. The corresponding figures for life insurance
were 13.9 %, 23.2 % and 46.2 % and 12.5%, 20.5% and 37.4% for non-life insurance. The book-value
of the assets held by German insurance companies in 1998 amounted 749.3 billion Euro, 61.7% of
which are held by life insurers, 8.4% by the pension funds and the funeral expense companies, 7.6% by
the health insurers, 12.1% by the non-life insurers, and 10.1% by the reinsurers. The market value of
the assets is much higher. Shares, investment funds as well as other fixed and non-fixed securities are
valued according to the so-called strict “lower of cost or market price principle”, meaning that current
assets have to be shown at cost value or market value at the balance sheet day or market value on any
previous balance sheet day, whichever is the lowest. This leads to the high amount of hidden reserves
of about 15.5% of the book value of the assets.
The Austrian Insurance Market
With a population of only about 8 million, Austria is a small country. This fact is also reflected in the
size of the insurance market. With a premium volume of 10.2 billion Euro in 1998, the Austrian
insurance market is less than 1/10 of the size of the German insurance market. The insurance sector
represents a nominal 5.4% of the country’s total GDP.
In 1998 there were 64 insurance companies licensed in Austria. According to the provisions of the
Austrian Regulatory Insurance Law, domestic insurers are allowed to conduct business only in the legal
form of a stock or in the form of a mutual company. Companies with headquarters outside the
European single insurance market that want to enter the Austrian market have to establish a branch
office. In 1998 branch offices represented only two foreign companies with headquarters outside the
European single insurance market in Austria. All insurance companies that are headquartered within
the European single insurance market are entitled to conduct business in Austria under the freedom to
provide services act. Comparable to the situation in Germany, only 0.3% of the total premiums of the
direct business is written under the freedom to provide services act.
Over the last decade, the development of the number of licensed insurers in Austria is rather stable.
Also the membership to the European single insurance market has not significantly changed the
number of market participants. Mostly due to the accession of Austria to the European Union, the
number of branch offices decreased constantly from 17 in the beginning of the 1980s to two in 1998.
Both were branches of insurers headquartered in Switzerland.
Most of the companies on the Austrian insurance market were licensed before the amendments to the
supervisory law that transformed the provisions of the Third Generation of Insurance Directives of the
European single insurance market into national law. Still these companies are entitled to offer
insurance products from more than one ”balance sheet division”2 under one legal entity to their clients.
Due to this fact just 6 insurance companies in Austria specialize in life insurance and 17 offer only non-
life insurance business other than health. However, in 1998 no company concentrated only on health
insurance. All the other companies were active in more than one balance sheet division. This fact has
to be taken into account when analyzing the annual statements of these insurance companies, because
due to the nature of the insurance business, distortions in the accountability of the data are possible.
Market concentration has increased over the last years. In 1998 the largest insurer in Austria held a
market share of 12.01 %. The three largest insurance companies controlled 31.53 % of the market and
2
Balance sheet divisions in Austria are health insurance, life insurance and non-life insurance other
than health insurance.
3
the 10 largest companies held 63.81%. The four largest insurance groups in Austria controlled a
market share of about 65%.
Like in Germany, concentration was the highest in the health insurance business, where the largest
company accounted for 30.3 % of the market. The largest three health insurers controlled 71.61 % of
the market, whereas the largest five health companies had a market share of 95.05 %.
The market share of the largest life insurance company was 12.02 %. The top three life insurance
companies held 31.88 %, the largest ten life insurance companies controlled 69.42 %.The largest
insurance company in the non-life business (other than health) held a market share of 15.26 %, the
largest three accounted for 38.49 % and the largest 10 controlled 75.23 %.
The life insurance business accounted for 40 %, health insurance for 11 % and non-life insurance other
than health insurance accounted for 49 % of the premium volume of 10.2 billion Euro in the primary
insurance business. The life insurance business plays an increasing role in private retirement savings,
due to the effects of the aging society on the social insurance system in Austria. Especially in the auto
and in the fire insurance we observe fierce competition, not started by companies doing business on
basis of the freedom to provide services act or the freedom of establishment act, but by domestic
insurers.
The book-value of the assets held by Austrian insurance companies was 43.5 billion Euro. Like in
Germany, the market value of the assets is higher, because certain components of the invested assets
have to be valued according to the so-called strict “lower of cost or market price principle”, meaning
that current assets have to be shown at cost value or market value at the balance sheet day or market
value on any previous balance sheet day, whichever is the lowest.
IV. DATA AND METHODOLOGY
The Data
To estimate efficiency, we obtained financial data for the German insurance companies directly from
either the balance sheet or the profit and loss account. We analyzed all life and all health insurers as
well as the largest 112 non-life insurance companies licensed in Germany, holding a market share of 74
% of the total reserves.
The data for the Austrian insurance market is obtained from the Austrian Insurance Statistics for 1998,
which is published by the Austrian Ministry of Finance3 (1999). The data set includes all insurance
companies, licensed in Austria and marketing their products on the Austrian insurance market. We do
not include the two branch offices of the Swiss insurance companies in Austria. Furthermore we do
include the data of the legal form of so called ”small mutual”. These insurers are “regional players”
having little influence on the Austrian insurance industry and operating under different technology,
which would bias our efficiency estimations. For the same reason we rejected data of pension funds,
funeral expense companies, foreign subsidiaries, and reinsurance companies.
Since most of the Austrian insurance companies offer more than one ”balance sheet division” under
one legal entity, we had to account for the distortions in the accountability of the data to one of the
three ”balance sheet divisions”. Since in Germany it is allowed to offer life insurance, health insurance
and non-life insurance business in the same group we also had to account for the same problems in the
accountability of the German data.
To estimate efficiency, we had to define firm inputs and outputs. A review of the literature shows no
clear consensus on the “best” input/output specification.4 In our definition of the inputs and the outputs
of insurance companies, we follow Cummins, Weiss and Zi (1999), Cummins, Turchetti and Weiss
(1996), and Cummins and Weiss (1993). For non-life insurance companies, we divide the output into
the risk-bearing function, the services related to insured losses, and into the intermediation function.
Following Cummins and Zi (1997), we divide the output of a life insurance company and of a health
3
In Austria and in Germany the Ministers of Finance serve as the Insurance Regulatory Authorities.
4
An overview on efficiency-research in the insurance industry is presented by Cummins and Weiss
1999.
4
insurance company into the risk-bearing function and the intermediation services. We proxy the risk
bearing function and the services related to insured losses in the non-life industry with the claims net of
reinsurance incurred in a certain year. The corresponding concept for the risk-bearing function in the
life insurance and in the health insurance is the incurred benefits net of reinsurance in a given year.
The intermediation function for non-life insurers is proxied with the total invested assets, for the life
and the health insurers with the total invested assets and the changes in reserves net of reinsurance.
To model the inputs of the insurance companies in this study, we use net operating expenses to proxy
the distribution of insurance products, the inputs of labor force, business services, and materials used in
the production of the insurance products. Equity capital and technical provisions net of reinsurance
proxy the inputs for the risk-bearing and risk-pooling function of the insurer. Summary statistics for
each of the three divisions for both countries are provided in Tables 1, 2 and 3.
Data Envelopment Analysis
Data envelopment analysis (DEA) is a linear programming technique that can be used to estimate the
efficiency of a firm or its management (referred to as decision-making units or DMU’s). Essentially,
all of the various versions of DEA models attempt to determine which of the N DMUs in the sample
determine the efficient frontier or the envelopment surface, with each DMU having m inputs and s
outputs. For DMU I, xiI is the ith input value and yrI is the rth output value with XI and YI representing
the respective input and output vectors used in the analysis. Surfaces can be of several types. In this
study we use a base model and construct a variable returns to scale (VRS) surface.
For the VRS surface, we solve a mathematical programming model for each of the DMUs in order to
define an efficient production frontier. In the programming model, we define a matrix of s outputs for
each DMU as Y and the matrix of m outputs for each of the DMUs as X. Assuming a non-oriented
model5 and a VRS efficient frontier, the mathematical model is as follows:
(1) VRS (YI,XI,u’,v’): min-(u’s+v’e); Yλ-s = YI;-Xλ-e=-XI;1λ=1;λ≥0;s≥0;e≥0,
where optimal values of the input and output variables for DMU I are denoted by the s-vector s’, the m-
vector e’, and the n-vector λ ’. Additionally, an optimal dual solution is given by the s-vector u’, and
m-vector v’, and the ώ vector that defines the intercept of the frontier in the VRS model.
Note that the vectors (u’,v’) define the specific lower bounds of the dual variables u and v referred to
as the relative prices. We use a standard or equal bounds technique such that when u’r = 1, r=1,…s and
v’i=1, i=1,….,m. Each of the N sets of values given by u’, v’, and ώ are the coefficients of the
hyperplane that define the surface of the efficient frontier. The values of u’r v’i are the prices. A DMU
is deemed efficient if it lies on the facet defining the hyperplane or envelopment surface. Note that for
the VRS model, the surface is defined as u’y-v’x + ώ = 0. The principle of evaluation for this type of
model is then given by the choice of the vectors u’ and v’.
The vectors (YI’,XI’) = (∑j=1,…n, λ’jYj, ∑j=1,…n λ’jXj) define a point on the frontier. If the DMU is
efficient, or lies on the frontier, λ’I = 1. For an inefficient unit, one that does not lie on the frontier,
(YI’,XI’) is referred to as the projected point. This point is a convex combination of the units that lie on
the envelopment surface for the VRS model. Total inefficiency is computed by the distance from the
actual point to the computed point. This is given by
∆’s = Y’I-YI for the output estimate of total inefficiency and
∆’e = XI-XI’ for the input estimate of total inefficiency.
The deviations from efficiency can be a result of failure to proportionally reduce inputs, failure to
proportionally augment outputs, and residual reductions and augmentations that can be accomplished
beyond the proportional changes. This can be expressed mathematically for outputs and inputs as
follows:
∆’s = (φ-1)YI+ψ’s
∆’e = (1-τ)XI+ ψ’e
5
For an overview see Charnes, Cooper, Lewin, and Seiford 1994.
5
where (φ-1) is the maximum possible proportional increase in outputs and (1-τ) is the maximum
proportional decrease in inputs, both with respect to the optimal projected point. Residual reductions
beyond proportional reductions are denoted by ψ.
In this study, we focus our results on the input and output efficiency measures in terms of efficiency
ratios. The two ratios are the input efficiency measure and the output efficiency measure, which are
defined as rates of change in relative prices. The input measure is denoted as i' and the output measure
is denoted by o’. For our VRS model, the measures are calculated as follows:
i’ = (U’YI + ώ)/v’XI
o’ = (v’XI- ώ)/u’YI
The DMU is efficient when i’ and o’ are equal to 1. Notice that i’ is constrained to be less than or
equal to 1 and o’ is constrained to be greater than or equal to 1.
V. RESULTS
Table 4 summarizes the efficiency results for each of the three sets of insurance companies in Austria
and Germany. For the input models, values of less than one indicate the firm could have reduced its
inputs and for the output models a result of greater than 1 indicates the firm could have augmented its
outputs. In particular, for the non-life input models, the average German firm needed to reduce inputs
by 17 percent to become efficient and the Austrian firm could have realized a 19 percent proportionally
reduction in input utilization. With respect to output efficiency for the non-life firms, again we find
that German firms operating closer to the efficient frontier with a mean efficiency score of 1.23
indicating that they needed to augment outputs proportionally 23 percent, while their Austrian
counterparts had a mean efficiency score of 1.41, indicating the need to augment output by 41 percent
to become efficient. The difference in output efficiency is significant as indicated by the p-value of
.016.
Turning to life insurance firms, the Austrian firms are significantly more efficient on the input side
with a mean efficiency score of .93 compared to a mean efficiency score for the German firms of .89.
However, this input efficiency advantage is offset by the German firms having a statistically significant
efficiency advantage on the output side. Hence, with respect to the life insurance firms, the evidence of
overall inefficiency is somewhat inconclusive.
In both countries, the health insurance firms are operating very close to their efficient frontier as
indicated by their mean efficiency measures near 1. The mean efficiency measures are not statistically
different.
Although a comparison of efficiency scores computed with different methods from different sample
groups that – when comparing different countries – in addition might be distorted by technology and
environmental factors might be misleading, the overall magnitude of the efficiency measures we found
in our study are in the range of efficiency scores for the insurance industry of prior studies. This is
especially true when we compare our results with the results of the study of Mahlberg and Url (1998)
for the Austrian market. Using a different input/output specification, different sample groups but the
same estimation technique, the authors found an efficiency score of .75 for the Austrian insurance
industry.6
The other prior study of the efficiency of the Austrian insurance market done by Anderson,
Ennsfellner, and Lewis (1999) used the same input/output specification, the same sample groups (for
Austria) but a different estimation technique as we do in this study. For the life insurance industry in
Austria, the mean input efficiency score was .87. The mean efficiency score for the non-life insurance
business was .77.7
6
Mahlberg and Url did not differentiate between non-life insurance companies, life insurance
companies and health insurance companies.
7
Anderson, Ennsfellner and Lewis did only differentiate between life and non-life insurance
companies.
6
In their study, Kessner and Polborn (1999) found an efficiency score for German life insurance
companies between .4 and .48 looking at a period from 1990 until 1993. However, in their study
Kessner and Polborn used the sum insured of newly written insurance policies and the sum insured of
the insurance pool as proxies for the output of the life insurance companies. Taking into account the
deregulation of the German regulatory system that took place in 1994 (i.e. their data are pre-
deregulation whilst ours are post-deregulation), it even seems to be more misleading to compare these
results to our results for Germany.
Finding significant output inefficiencies in the Austrian non-life insurance market as compared to the
German insurance market might in part be explained with the high concentration of the market.
However, we might assume under the freedom of establishment act and under the freedom to provide
services act competition should be healthy. Moreover, the fact that little entry and exit has existed in
Austria provides evidence that the market is relatively competitive. With relative ease of entry, we
have not seen many firms moving to transact business in Austria. If firms were earning positive
economic profits we would expect growth in the number of firms in the market. In Austria, we see a
few large firms dominating the industry. If scale economies are present this too could represent an
efficient situation. But it may be the case that these few large firms are dominating the market and
essentially creating barriers to entry, thus preserving positive economic profits. In this setting,
consumers would suffer, as the firms would be able to provide lower quality products at a higher cost.
But also the size of the market could explain these inefficiencies. The Austrian insurance market is a
small market as compared to the German market. Given that the size of the market influences also the
size of the companies it can be expected that the average Austrian insurance company is smaller than
its German counterparts thus creating disadvantages in a scale and in a scope economies sense. In
addition, the small size of the market might be a factor that has made the Austrian market a relatively
unattractive location for new firms.
Input efficiency of the Austrian life insurance companies might in part be explained by the fact that in
Austria life insurers with a market share of about 25 % sell its products almost exclusively via banks.
One reason for this result could be that bancassurance might influence efficiency in a positive way.
We strongly believe that other factors of the insurance industry do not provide reasonable explanations
for the observed differences in the results. Austria and Germany share about the same regulatory
system, the industries show a comparable insurance culture, and similar services are provided in
comparable proportions. In addition, insurance technology defined as a set of specific methods that
insurance companies use to generate a certain amount of output with a certain set of inputs is also about
comparable in the two markets. However, we hope that future steps in our research process show more
detailed explanations for the observed differences.
Now it also seems important to know which of the inputs and outputs contribute the most to the total
inefficiency of the firms not operating on the efficient frontier. With this information, the managers of
the inefficient firms can pinpoint where they need to enhance output and where they should be able to
reduce inputs in order to operate in a manner competitive with the better performing peer group
members. In order to do this, we sorted out all of the inefficient firms and computed their total
deviations from the efficient frontier and then computed the percentage of that total deviation that was
a function of each of the inputs and outputs that define their respective production functions. The
results of these computations are illustrated in Table 5.
With the exception of the German health insurance industry, in all other categories the most important
reason for the inefficiencies is the failure of the companies to increase the output of the total invested
assets. This is especially true for the life insurance companies. While in Germany about 74 % of the
inefficiencies in the life insurance business can be explained with the output of total invested assets,
this output accounts for more than 81% of the inefficiencies of Austrian life insurers. The major reason
why our results show inefficiencies in the total invested assets in both countries seems to be the fact
that in both countries major parts of the total invested assets have to be valued according to the so-
called strict “lower of cost or market price principle”. This principle states that current assets have to
be shown at cost value or market value at the balance sheet day or at market value on any previous
balance sheet days, whichever is the lowest. Therefore the market value of the invested assets is higher
than shown in the balance sheet, indicating a higher output of total invested assets. Especially the life
insurance results, where we found that the evidence of overall inefficiency is inconclusive, have to be
7
explained very carefully taking into account these distortions. The second largest factor for
inefficiencies of the German life insurance firms is the failure to increase changes of reserves net of
reinsurance that accounts for about 14 % of the problem. The second largest reason for inefficiencies
of Austrian life insurers is the use of the input “technical provisions net of reinsurance” which explains
about 11 % of the inefficiencies. The fact that a life insurance company keeps a high level of
provisions may not be the consequence of the excessive use of an input. On the contrary, it could be
argued that the “inefficient” use of the input “technical provisions net of reinsurance” could be a
prudent response to the need of the customers for a certain level of security.
In the German and in the Austrian non-life insurance business, about 58 % of the inefficiencies can be
explained by the failure to augment total invested assets. Differences between the two countries can be
identified when we look at the output “claims incurred”. In Austria, claims incurred contribute 34 % to
the total inefficiency of non-life insurers, while in Germany this figure is about 14 %. In Germany the
non-life insurance companies’ use of operating expenses net of reinsurance highly contributes to their
inefficiency. About 22 % of the inefficiencies of inefficient German non-life insurers can be explained
by this input, leading to the question for the reason for this problem.
We found that health insurance companies in both countries are operating very close to their efficient
frontier. However, the inefficient health insurance companies in Austria suffer from an inefficient use
of total invested assets, where the valuation of this output seems to be the reason for this problem.
Contributing about 30 % to the inefficiencies of Austrian health insurance companies, the incurred
benefits net of reinsurance are the second largest problem. Operating expenses net of reinsurance and
equity capital are also aspects that each are responsible for about 12 % of the problems of the
inefficient health insurance companies in Austria. In Germany the output of the health insurers seems
to be the major reason for inefficiencies of the health insurance business. Above all, the failure to
produce enough of the output “changes in reserves net of reinsurance” (53 %) contributes to the
problems of the inefficient German health insurers, followed by total invested assets (18%) and
incurred benefits net of reinsurance (15 %).
VI. CONCLUDING REMARKS
In this paper we compared input and output efficiencies between the German and the Austrian
insurance market. Using DEA, we found significant differences in the efficiencies of the non-life
output between the two countries. While German non-life insurance companies should increase output
proportionally 23 % to become efficient, their Austrian counterparts need to augment output by 43%.
Input efficiency of Austrian life insurance companies is significantly higher than that of German life
insurers, whereas output efficiency of German life insurer is significantly better than that of life
insurance companies in Austria. Health insurers in both countries work close to their efficient frontier,
observed differences in the efficiency are not significant.
In both, the German and the Austrian insurance market, the failure of insurance companies to increase
the output of total invested assets contributes largely to the inefficiencies. Provisions in the relevant
laws that determine how to value insurers’ assets might be a reason for these results. This is especially
true for the life insurance industries where, due to the nature of the business, long-term investments
play a major role. However, we identified differences in the contribution of inputs and outputs to the
inefficiencies of the insurance companies in our set of countries. While in Germany the second largest
reason for inefficiencies of life insurance companies is the failure to increase changes of reserves net of
reinsurance, second important for Austria is the inefficient use of technical provisions net of
reinsurance. Operating expenses net of reinsurance highly contributes to the German non-life insurers,
while in Austria claims incurred is a more important reason for the inefficiencies of non-life insurance
companies. Different to Germany, total invested assets contribute the most to the inefficiencies of the
Austrian health insurance companies, followed by incurred benefits, the inefficient use of equity capital
and problems with the operating expenses net of reinsurance. In Germany, the failure to produce
enough of the output “changes in the reserves net of reinsurance” is the biggest problem for the
inefficient health insurance companies.
Our future research has to determine the reason for the differences in efficiency between the German
and the Austrian insurance market found in this study. However, we think that for this study, the
assumption of about comparable environmental conditions for these insurance markets is applicable.
Therefore it seems appropriate to pool all insurance companies of our set of countries and to build a
common frontier. Under these assumptions, the common frontier is based on the belief that the
differences in efficiency are primarily a function of managerial decisions. However, also country-
8
specific differences could be a reason for observed differences in efficiency scores between our set of
countries. Although we do not expect too many changes in our results, in a next step of our research
we will reject the assumption of comparable environmental conditions for the German and the Austrian
insurance market and introduce country specific environmental variables in our frontier estimations.
The importance of the introduction of environmental variables even will increase if we extend our
cross-country comparison to other countries of the European Union. Following the study of cross-
country comparison of banking systems presented by Dietsch and Lozano-Vivas (2000) we would like
to introduce the following three categories of environmental variables in our frontier estimations:
Variables that describe the main macroeconomic conditions that determine the demand for insurance
products;
Variables that describe the structure and regulation of the insurance industry;
Variables that describe the accessibility of insurance services.
We expect that the introduction of these variables might help to explain cross-country differences in the
efficiency of insurance companies in greater detail. These results could provide a sound basis for
policy makers and insurance managers to identify practices that would promote higher efficiency in the
insurance markets, thus helping customers to reap the rewards associated with higher efficiency.
9
APPENDIX
Table 1a: Summary Statistics
Life Insurance Firms in Austria (1998)
Input 1 Input 2 Input 3 Output 1 Output 2 Output 3
Mean 16669 26617 684090 94567 31903 746055
Median 10096 10476 314036 39144 13991 336841
Standard 20483 39033 906308 128158 59185 993826
N= 40
Table 1b: Summary Statistics
Life Insurance Firms in Germany (1998)
Input 1 Input 2 Input 3 Output 1 Output 2 Output 3
Mean 1634653 1623047 512562 1981339 14932 1108273
Median 25024 21999 1017327 116852 14932 1108273
Standard 104849 96833 7288821 838917 539324 8199976
N= 117
Input 1= operating expenses net of reinsurance
Input 2 = equity capital
Input 3 = technical provisions net of reinsurance
Output 1 = incurred benefits net of reinsurance
Output 2 = changes in reserves net of reinsurance
Output 3 = total invested assets
10
Table 2a: Summary Statistics
Health Insurance Firms in Austria
Input 1 Input 2 Input 3 Output 1 Output 2 Output 3
Mean 19086 30817 200533 123142 15370 268735
Median 15609 11264 171916 110071 15261 188994
Standard 80658 47899 224851 119393 14229 319575
N= 9
Table 2b: Summary Statistics
Health Insurance Firms in Germany
Input 1 Input 2 Input 3 Output 1 Output 2 Output 3
Mean 57835 63059 1334438 379147 1155 1389714
Median 35024 40903 522613 174609 2714 537566
Standard 73879 88195 2064063 545714 259055 2172565
N= 42
Input 1= operating expenses net of reinsurance
Input 2 = equity capital
Input 3 = technical provisions net of reinsurance
Output 1 = incurred benefits net of reinsurance
Output 2 = changes in reserves net of reinsurance
Output 3 = total invested assets
11
Table 3a: Summary Statistics
Non-Life Insurance Firms in Austria
Input 1 Input 2 Input 3 Output 1 Output 2
Mean 27269 54498 24459 148159 191946
Median 6658 11617 3470 27647 57339
Standard 49081 96171 54607 276856 321883
N= 55
Table 3b: Summary Statistics
Non-Life Insurance Firms in Germany
Input 1 Input 2 Input 3 Output 1 Output 2
Mean 86450 251003 476440 233257 776540
Median 45267 102724 202169 107604 319865
Standard 129659 352152 879998 352171 1302265
N= 112
Input 1= operating expenses net of reinsurance
Input 2 = equity capital
Input 3 = technical provisions net of reinsurance
Output 1 = claims incurred net of reinsurance
Output 2 = total invested assets
12
Table 4: Mean Efficiency Differences
Efficiency German Mean Austrian Mean P-Value from T-Test
Measure Assuming Unequal
Variance
Non-Life Input .83 .81 .55
Non-Life Output 1.23 1.41 .016**
Life Input .89 .93 .037**
Life Output 1.03 1.07 .059*
Health Input .95 .92 .58
Health Output 1.02 1.12 .28
Table 5: Total Inefficiencies and Percentage Decomposition
Sample Set % i’1 % i’2 % i’3 % o’1 % o’2 % o’3
German 0.22601 0.053026 0 0.138544 0.582409 NA
Non-Life
Austrian 0.03119 0.0086931 0.036457 0.3442218 0.5794342 NA
Non-Life
German 0.04334 0.015846 0.02925 0.033566 0.137515 0.740476
Life
Austrian 0.01247 0.02021 0.11297 0.029423 0.011106 0.813811
Life
German 0.08170 0.05452 0 0.152871 0.52519 0.1857
Health
Austrian 0.11813 0.12819 0 0.3067 0.030447 0.416527
Health
For non-life insurance:
Input 1 (I1)= operating expenses net of reinsurance
input 2 (I2)= equity capital
input 3 (I3)= technical provisions net of reinsurance
output 1 (O1)= claims incurred net of reinsurance
output 2 (O2)= total invested assets
For life and health insurance:
Input 1 (I1)= operating expenses net of reinsurance
Input 2 (I2)= equity capital
Input 3 (I3)= technical provisions net of reinsurance
Output 1 (O1)= incurred benefits net of reinsurance
Output 2 (O2)= changes in reserves net of reinsurance
Output 3 (O3)= total invested assets
13
REFERENCES
Allen, L., Rai, A., “Operational Efficiency in Banking: An International Comparison.” Journal of
Banking and Finance, 20, 1996, 655 – 672.
Anderson, R.I., Ennsfellner, K.C., Lewis, D., “Production Efficiency in the Austrian Insurance
Industry: A Bayesian Examination” [under review].
Austrian Ministry of Finance, “Austrian Insurance Statistic for the Year 1998.” Vienna 1999.
Berg, S., Bukh, P., Forsund, F., “Banking Efficiency in the Nordic Countries: A Four-Country
Malmquist index analysis.” Working Paper, University of Arhus, Denmark, 1995.
Berg, S., Forsund, F., Hjalmarsson, L., Suominen, M., “Banking Efficiency in the Nordic Countries.”,
Journal of Banking and Finance 17, 1993, 371 – 388.
Bergendahl, G., “DEA and Benchmarks for Nordic Banks.” Working Paper, University of
Gothenburg, Sweden, 1995.
Berger, A.N., Cummins, J.D., Weiss, M.A., ”The Coexistence of Multiple Distribution Systems for
Financial Services: The Case of Property-Liability Insurance.” Finance and Economics Discussion
Series, no. 95/22, Washington D.C., 1995.
Berger, A.N., Demsetz, R.S., Strahan, P.E., “The Consolidation of the Financial Services Industry:
Causes, Consequences and Implications for the Future.” Journal of Banking and Finance, 23, 1999,
135 – 194.
Berger, A.N., Humphrey, D.B., “Efficiency of Financial Institutions: International Survey and
Directions for Future Research.” European Journal of Operational Research, 98, 1997, 175 – 221.
Berger, A.N., Humphrey, D.B., “Measurement and Efficiency Issues in Commercial Banking.” In:
Griliches, Z. (Ed.), “Output Measurement in the Service Sector.” Chicago, London 1992, 245 – 296.
Berger, A.N., Hunter, W.C., Timme, S.G., “The Efficiency of Financial Institutions: A Review and
Preview of Research - Past, Present, and Future”, Journal of Banking and Finance, 17, 1993, 221 –
249.
Born, P., Gentry, W., Viscusi, W.K., Zeckhauser, R.J., “Organizational Form and Insurance Company
Performance: Stocks versus Mutuals.” NBER Working Paper Series, National Bureau of Economic
Research, Working Paper no. 5246, September 1995.
Bundesaufsichtsamt, “Geschäftsbericht des Bundesaufsichtsamtes für das Versicherungswesen 1998,
Teil B [Annual Report of the Bundesaufsichtsamt for the Insurance Industry 1998, Part B].” Berlin
1999.
Charnes, A., Cooper W.W., Lewin, A.Y., Seiford L.M., “Data Envolopment Analysis: Theory,
Methodology, and Application.” Boston, Dordrecht, London, 1994.
Cummins, J.D., Tennyson, S., Weiss, M.A., “Consolidation and Efficiency in the US Life Insurance
Industry.” Journal of Banking and Finance, 23, 1999, 325 – 357.
Cummins, J.D., Turchetti, G., Weiss, M.A., ”Productivity and Technical Efficiency in the Italian
Insurance Industry.” Working Paper, Wharton School of the University of Pennsylvania, no. 96-10,
1996.
Cummins, J.D., Weiss, M.A., “Analyzing Firm Performance in the Insurance Industry Using Frontier
Efficiency and Productivity Methods.” In: Dionne, G. (Ed.), “Handbook of Insurance Economics.”
Boston, 2000.
Cummins, J.D., Weiss, M.A., “Measuring Cost Efficiency in the Property-Liability Insurance
Industry.” Journal of Banking and Finance, 17, 1993, 463 – 481.
14
Cummins, J.D., Weiss, M.A., Zi, H. 1999, „Organizational Form and Efficiency: An Analysis of Stock
and Mutual Property-Liability Insurers," 1999, Management Science, 45, 1999, 1254 – 1269.
Cummins, J.D., Zi, H., “Measuring Cost Efficiency in the US Life Insurance Industry: Econometric
and Mathematical Programming Approaches.” Working Paper, Wharton School of the University of
Pennsylvania, no. 97-03, 1997.
Dietsch, M., Lozano-Vivas, A., “How the Environment Determines Banking Efficiency: A Comparison
between French and Spanish Industries.” Journal of Banking and Finance, 24, 2000, 985 – 1004.
Fecher, F., Kessler, D., Perelman, S., Pestieau, P., “Productive Performance of the French Insurance
Industry.” The Journal of Productivity Analysis, 4, 1993, 77 – 93.
Fecher, F., Pestieau, P., “Efficiency and Competition in O.E.C.D. Financial Services.” In: Fried, H.O.,
Lovell, C.A.K., Schmidt, S.S. (Eds.), “The Measurement of Productive Efficiency – Techniques and
Applications.” Oxford 1993, 374 – 385.
Ferrier, G.D., “Comment on Hughes, Lang, Mester, and Moon and Cummins, Tennyson, and Weiss.”
Journal of Banking and Finance, 23, 1999, 358 – 366.
Gardner, L.A, Grace, M.F., “X-Efficiency in the US Life Insurance Industry.” Journal of Banking and
Finance, 17, 1993, 497 – 410.
Kessner, E., Polborn, M., “Eine Effizienzanalyse der deutschen Lebensversicherer – die Best Practice
Methode [Efficiency Analysis of German Life Insurers – the Best Practice Method].” Zeitschrift für die
gesamte Versicherungswissenschaft, 1999, 469 – 488.
Mahlberg, B., Url, T., “Effects of the Single Market on the Austrian Insurance Industry.” Working
Paper, presented at the Annual Meeting of the Geneva Association in Vienna, September 1998.
Pastor, J.M., Pérez, F., Quesada, J., “Efficiency Analysis in Banking Firms: An International
Comparison.” European Journal of Operational Research, 98, 1997, 395 – 407.
Standard & Poor’s, “Germany – Insurance Market Profile.” 1998.
Yuengert, A.M., “The Measurement of Efficiency in Life Insurance: Estimates of a Mixed Normal-
Gamma Error Model.” Journal of Banking and Finance, 17, 1993, 483 – 496.
15
