Masteruppsats i offentlig förvaltning [VT/2012]
Förvaltningshögskolan, Göteborgs universitet Seth Sander 870714-4930
Handledare: Lena Lindgren Examinator: Patrik Zapata
Efficient Privatization of SOEs –
The Importance of Operationalization in
Policy Evaluation.
Abstract
In the beginning of this thesis a problem regarding efficiency operationalization in previous studies of SOE privatization was introduced. The purpose of the thesis was then established as discerning in what ways various operationalizations of efficiency influenced the obtained results and their validity.
After a brief introduction to the history of privatizations and the progression of ideas and
feelings regarding privatization previous empirical studies of privatization were detailed. It
was established that these previously conducted empirical studies of privatization were quite
similar in terms of their operationalizations of efficiency as well as their end results. From
these previous studies an operationalization was derived based on their similarities, which was
labeled operational efficiency. I then argued that these previous empirical studies were in fact
evaluations based on the definition of evaluation supplied by Vedung. From the basis of
evaluation theory I proceeded to question the operationalizations in the detailed evaluations of
privatization and, by extension, the validity of their conclusions. The basis of the criticism of
the operationalizations in the detailed evaluations of privatization was based primarily on the
lack of adherence to goals. In order to identify what the potential consequences of the
operationalizations in the detailed privatization evaluations could be two analyses of a recent
case, Apoteket AB, were carried out. One of the analyses was carried through the
identification and operationalization of the goals of the privatization; this process was labeled
goal fulfillment. The other was carried out using the previously defined operational
efficiency. The results of these two analyses were then compared. Despite the purposes of the
analyses being the same the results provided information regarding completely different
aspects of the privatization and the connection between operational efficiency and goal
fulfillment was nonexistent. I then argued that the construct validity of the measurement
operational efficiency was sorely lacking and that by extension the conclusions regarding the
efficiency of privatization as a policy tool are questionable.
Abstract...2
Introduction...4
What is the purpose of this paper?... 4
Important Definitions and Distinctions... 6
What is an SOE?... 6
Can Apoteket AB be considered an SOE?...7
What is a privatization?... 7
Can the reregulation of the Swedish pharmaceuticals monopoly be considered a privatization?...8
Outline of the Thesis...9
A Short Historical Background of Privatization...11
Privatization and Efficiency, Defining Operational Efficiency and Goal Fulfillment...14
Empirical Studies of Privatization... 14
An Alternative View of Efficiency... 18
What, besides goal fulfillment, is needed in order to assess efficiency?...27
Methodology – Measuring the Effects of Privatization, Dual Approaches...28
Sources and Collection of Data... 31
Data used to measure changes in operational efficiency...32
Empirical Results, Goal Fulfillment – Pre- and Post-Privatization Data...34
Empirical Results – Pre- and post-privatization operational efficiency...47
Conclusions...52
Reflections...56
Thoughts regarding my choice of case... 56
The need for further research... 56
Summary...58
References...59
Printed Sources... 59
Public Sources, Annual Reports and other economic data...60
Web Based Sources... 62
Introduction
What is the purpose of this paper?
Privatization is a relatively new phenomenon often said to have started with the Thatcher government in Britain.
1Since then privatization has become widespread, primarily but not exclusively, in the industrialized world.
2Sweden has not been an exception to this, somewhat surprisingly considering a history of nearly unbroken rule by a social democratic political majority.
Over the years a multitude of empirical studies regarding privatization of state owned enterprises (SOEs) have been conducted. In much of the recent empirical analysis of privatization efficiency has been assessed through various comparisons of SOE performance before and after a privatization.
3As I will show later in this thesis the pre- and post- privatization performance of former SOEs is generally measured through a number of key indicators of SOE efficiency. The more comprehensive studies repeat this process for a large number of former SOEs in different countries. An overview of these empirical studies of newly privatized SOEs shows very convincing evidence for the benefits of privatization in terms of increased efficiency and output of SOEs.
4While these results are quite convincing and clear-cut I find that there is an aspect of this empirical research that warrants further investigation and it has to do with the way the term efficiency is operationalized in these previous studies. Closer examination of the term efficiency reveals that the factors taken in to account in the previous studies may be insufficient.
5Vedung for example describes efficiency in terms of “The programs’ capacity to produce predicted outcomes that fulfill the goals of the program (…)”.
6I do not believe that the operationalizations present in the previous studies account for goal fulfillment as an efficiency criterion and it is this problem that I have studied.
1
Megginson et.al (1998) p. 562
Dyck (2001)3
For a more detailed description see section 4 starting at page 13 of this thesis.A definition of what does and does not constitute an SOE is provided in the following section.4
Jordahl (2008)5
For a comprehensive discusison of this see pages 14 – 19 of this thesis.6
Vedung (2009) p. 31By analyzing a recent privatization through the use of the conventional operationalization and methodology found in previous studies as well as an alternative, theoretically derived, operationalization and methodology I will attempt to discern in what ways these operationalizations of efficiency can influence the obtained results and their validity. The central question guiding this thesis will therefore be: How does the operationalization of efficiency in the detailed previous empirical studies of privatization affect results?
In order to answer this question I will:
A: Analyze previous empirical studies of privatization in order to identify an established operationalization and methodology.
B: Theoretically derive an alternative operationalization and methodology.
C: Analyze a recent case of privatization using both operationalizations and methodologies separately in order to be able to compare the results to one another and analyze how results are affected by the disparate operationalizations.
Any research that can increase the general understanding of how the effects of privatizations of SOEs are affected by evaluation-methodology would be worthwhile and could potentially improve the quality of political decision-making on a national and supranational level. But perhaps more importantly this thesis has the potential to create awareness of the need for more interdisciplinary scientific analysis of privatizations as opposed to analyzing the economic and sociopolitical effects independently. In fact, large-scale privatizations are still being planned, for example in Sweden
7and in the interest of improving these future decisions more research on the subject of privatization is crucial.
To this end I am going to examine the most recent privatization in Sweden, namely the state owned pharmaceutical retail monopoly Apoteket AB. Apoteket AB was, until July 2009, the sole provider of retail pharmaceuticals in Sweden. In 2009 the market was reregulated and a large section of Apoteket’s assets were sold. The privatization process will be described in more detail in a later section. The primary reasons for choosing this particular case were its size and the fact that it was recent. Completion of the thesis requires effects of the
privatization to be measured and at the time I reasoned that these effects may be more evident
7
Odell (2009)for a large privatization as opposed to a smaller one. The fact that the privatization was quite recent meant that it had yet to be evaluated by any national government agency and I also assumed that data regarding the privatization would be abundant.
Before the examination of this case can begin there are some important definitions and distinctions that need to be made.
Important Definitions and Distinctions
A definition of the term efficiency is obviously very important for this thesis and such a definition will be forthcoming. Unfortunately placing a definition of efficiency in this section interfered with flow of the thesis as a whole and it is therefore part of a later section.
Another important questions is likely whether my chosen case, Apoteket AB, is even an example of an SOE privatization. This question has at least two parts:
1. Can Apoteket AB prior to the reregulation of the retail pharmaceuticals market be considered an SOE?
2. Was the reregulation actually a case of privatization?
Each of these questions in turn need to be answered in at least two parts:
1a. What is an SOE?
1b. Can Apoteket AB be considered an SOE?
2a. What is a privatization?
2b. Does the reregulation of the Swedish pharmaceuticals monopoly fall under this definition of privatization?
What is an SOE?
1a: There are two defining features of an SOE. The first is that the nation state is the
controlling owner of the enterprise and the second is that goods or services are sold.
8The
important distinction to make here, especially in the Swedish context, is between an SOE and
a state owned organization that provides a service but is primarily financed indirectly through
taxes. In the Swedish case a school or hospital that provides the overwhelming majority of
their services essentially free of direct charge are not to be considered SOEs as opposed to for
example formerly state owned Svensk Stål AB (SSAB) a steel/mining company which sold its
8
The World Bank(1995) p. 26goods and services both prior to and after privatization. This is an important distinction to make, as nearly every part of the organizational structure, as well as the goals and general culture of these types of organizations is likely to be entirely different. Thus it stands to reason that the theories explaining the privatization of one do not necessarily apply to the other. This may well be an interesting subject for future research to examine but privatizations of non-SOEs are not covered within the scope of this thesis.
Can Apoteket AB be considered an SOE?
1b: The primary purpose of the company was to fulfill the role as sole vendor of
pharmaceuticals in Sweden. While Apoteket AB did in fact provide certain free public service functions such as the poison information hotline, (giftinformationscentralen) these services were not part of the privatization and are still provided in a public service capacity by Apoteket AB after the reregulation.
9The privatization encompassed the business side of Apoteket AB which sold pharmaceuticals to consumers in order to finance its own existence.
Therefore this criterion is fulfilled.
What is a privatization?
2a: The defining feature of a privatization is the relinquishment of state control of an organization in favor of private actors, however this definition is oversimplified. In fact, in order for an SOE to be considered privatized there is no need for complete liquidation of the state’s stake; this is generally referred to as a partial privatization where part of the ownership is sold but the state retains voting control. A full privatization is often handled through a sale of stock in the SOE to the point where the state is no longer the controlling owner.
10A sale of an SOE to another state or other company controlled by the state is generally not considered a privatization.
11In the scientific literature the term privatization is occasionally supplanted by other terms such as divestiture. In this thesis I have chosen to use the term privatization as I believe it to be the most widely used.
Can the reregulation of the Swedish pharmaceuticals monopoly be considered a privatization?
2b, What differentiates the case of Apoteket AB from the most common types of privatization is that the enterprise Apoteket AB was broken up in to two main categories; one that is still under state ownership and one that was broken up into clusters which in turn were sold to 9
Prop. 2008/09:145 (2009)10
For a discussion of what constitutes controlling ownership see La Porta et.al (1999)11
Megginson et.al (1998) p. 62 Jordahl (2008) p. 9.private business. Therefore while approximately 60% of Apoteket AB’s assets were sold, the voting rights and the ownership of the brand name Apoteket AB, along with the remainder of the assets are still in the hands of the state.
12The 60% of Apoteket AB’s assets that were sold to private investors now directly compete with the remainder of the SOE in a mostly open market. Considering that there was significant transfer of ownership from public to private the reregulation can be considered a privatization even though what is left of the SOE is still under state ownership. A privatization of 60% of an SOEs assets is a clear example of a privatization and therefore this criterion is fulfilled.
12
Prop 2008/09:145 (2009) p. 84fOutline of the Thesis
The historical background that follows this section serves to put the current theoretical and empirical studies of privatization into a historical context and describe the process which has led to where the field is today.
The following theoretical section will begin with a brief overview of the privatization literature. This overview is not to be considered all-encompassing but serves the purpose of giving the reader an introduction to the central ideas of the empirical work that has been conducted on privatizations of SOEs. I will then move on to discuss what I consider methodological shortfalls in the empirical study of privatization and the contribution I wish to make to the general understanding of the privatization phenomenon and its evaluation. The methodological shortfalls in the previous literature pertain to the definition of efficiency and I will argue that the narrow definition of efficiency used in many empirical studies does not effectively assess the effects of an SOE privatization on a society. I will then present an alternative way of assessing the efficiency of a privatization.
In the theoretical section I give an account of a number of studies of privatizations and identify the parameters used to measure the efficiency of a privatization effort. I also argue that this way of assessing the efficiency of privatizations may not be optimal and propose an alternative way of measuring the efficiency of privatizations mostly by supplanting operational efficiency with goal fulfillment. I also identify the general goals of my chosen case as increased consumer access and lower prices of pharmaceuticals. In order to assess these goals more easily I have formulated a number of hypotheses pertaining to each goal.
These hypotheses will later be tested and the results of the hypotheses tests will be used as a basis for an assessment of goal fulfillment. The methodological section presents these hypotheses and defines how they will be tested and what data will be used. It also more clearly defines how the assessment using the established efficiency definition will be carried out.
The empirical section summarizes the data gathered and tests the hypotheses based on the
data from before and after the privatization of Apoteket AB. The results of these tests, as well
as their implications for fulfillment of the privatization’s goals are then analyzed in a short
analysis section. The pre- and post-privatization data points for the theoretically derived performance indicators are also presented and synthesized.
In the conclusions section the results of the goal fulfillment assessment will then be compared
to the results of the analysis of operational efficiency and analyzed using the theoretical tools
developed earlier in the thesis.
A Short Historical Background of Privatization
Privatization is often identified in political discourse as pertaining to a certain ideology or political standpoint such as, but not limited to, neo-liberalism
13but I believe this view can be misleading. In the past few decades, governments from all parts of the political spectrum have been carrying out privatizations of various SOEs.
14Instead I believe a great deal more can be understood if privatization is viewed as a phenomenon not exclusive to governments with a certain political tradition, i.e. conservative, socialist or liberal, but rather as a paradigm shift from the use of one form of policy too to another. What is considered favorable ownership structure changes over time and what is popular one day may not be so the next; and as a matter of definition, widespread privatization must have been preceded by widespread state ownership or there would be nothing to privatize.
Shirley & Walsh describe how state ownership experienced surging popularity during the 1930s, 1940s and 1950s among many developed nations and somewhat later among developing nations. At the time this was seen as the cure for market failures, which were considered the cause of all manner of problems and economic injustices. In parts of the developing world the rise of SOEs was also brandished as a means of state economic independence and a move away from foreign meddling in the economy. Although there was some theoretical criticism of state ownership at the time it did not have any significant impact until the 1960s and 70s.
15During the 50s and 60s the attitudes toward SOEs in the developed world started to change.
While the idea of state ownership was still generally looked upon favorably there was increasing concern over poor performance and inefficiency of SOEs. Nellis describes how the general mindset regarding SOEs changed and how the World Bank, among other organizations started to credited governments for the purpose of reforming SOEs:
“The Bank had assisted member governments in this field in hope that, in addition to providing basic and needed services in an efficient and cost-effective manner (…) public
13
Alliansfritt Sverige (2012)Corpwatch (2012) Socialistalternative (2012) The World Bank (1995) p. 4
14
Megginson et.al (1998) p. 4415
Shirley & Walsh, (2001) p. 3enterprises would assist the development of strategic sectors, gain access to commercial credit that would be denied to small private businesses, fill entrepreneurial gaps, empower numerically large but economically weak segments of the population, maintain employment levels, and raise the level of savings and investment.”16
Nellis goes on to explain that by the end of the 70s there were considerable doubts as to whether these goals were achievable through reform. Despite repeated reforms at great expense the reformed SOEs failed to meet the reform goals. Most SOEs continued to perform as inefficiently as before and the performance of some actually worsened. In some developing countries the burden of lossmaking SOEs overwhelmed budgets and banking systems. The economic turmoil even created problems for profitable firms, which could not raise sufficient investment capital.
17These realizations initiated a paradigm shift at the World Bank and among researchers: “By the mid-1980s, the dominant view among Bank staff working on SOE reform was that the available set of improvement mechanisms, short of ownership change, was hard to apply and even more difficult to sustain, and that more hope was offered by privatization”.
18The conservative British government led by Margaret Thatcher in the late 1970s coined the term privatization to describe the transfer of ownership from public to private and this is often thought of as the start of a privatization paradigm.
19The privatizations of the Thatcher government were not the first privatizations of that era (both West Germany and Chile had sold SOEs prior to that). However the impact of the wave of privatizations in Britain in the early 1980s was much greater.
20The privatizations carried out by the Thatcher government were quite controversial at the time. There was substantial political resistance from opposition politicians to privatizations and even threats of re-nationalization by the Labour party.
21Despite political opposition the initial British privatizations were well-received and Margaret Thatcher secured a second term in office in 1983. The apparent success of the British privatization effort sparked interest
16
Nellis (2002) p. 2f17
Nellis (2002) p. 318
Nellis (2002) p. 3f19
Shirley & Walsh (2001) p. 3 Megginson et.al (1998) p. 4420
Jordahl (2008) p. 1121
Megginson et.al (1998) p. 47among other western European countries and similar programs of privatization were undertaken. In France alone 22 large SOEs were sold during 1986-87.
22The same period also saw major privatizations in countries such as Japan and the United States. The privatization of the Japanese SOE Nippon Telephone and Telegraph (NTT) was by far the largest ever undertaken at the time, worth approximately USD 77 billion. In the United States the most notable privatization was that of Conrail and it, while much smaller than the privatization of NTT, received significant publicity. Both of these major privatizations were well-received and generally deemed quite successful.
23These early privatizations did not only draw significant political interest but scientific as well, but most of the empirical findings generally lacked statistical significance and the results of these studies were mostly inconclusive.
24More recent empirical work (which will be discussed at greater length in a later section) has provided less ambiguous results. Megginson et.al stipulate that:
“More recent theoretical and empirical studies have offered stronger support for the dual propositions that private firms outperform SOEs and that privatization itself increases the operating efficiency of divested firms.”25
As privatizations have continued to change ownership structures in countries around the world the privatization paradigm has become quite influential, in no small part due to the abundance of empirical research endorsing it.
2622
Jordahl (2008) p. 1123
Megginson et.al (1998) p. 58f24
Megginson et.al (1998) p. 4525
Megginson et.al (1998) p. 4626
Dyck (2001) p. 59Privatization and Efficiency, Defining Operational Efficiency and Goal Fulfillment.
Empirical Studies of Privatization
Privatizations, public/private partnerships and other collaborations have become very popular during the last three decades. This process has been described as a paradigm shift from a public/private relationship that was seen in terms of clearly defined lines and clear hierarchical exertions of power from top to bottom to a “blurring of the lines” between public and private.
27This paradigm shift has spawned a vast amount of scientific literature in areas such as governance, privatization and new public management. The apparent success of many of these privatizations and partnerships in increasing operational efficiency and solving certain problems has helped spread these ideas around the world through organizations such as the World Bank.
The empirical literature concerning the matter of public versus private ownership can be divided in to two categories by way of methodology.
28The first category, generally made up of earlier studies, compared the performance of SOEs to that of private enterprises in comparable fields and over time. This approach generally suffers from the methodological problem of differentiating the ownership effects from other differences that invariably exist between different enterprises.
29The World Bank explains the second category thus:
“More recent work compares the performance of SOEs before and after divestiture, divested with undivested SOEs, or divested firms with a hypothetical counterfactual in which the same firm is assumed to continue under state ownership.”30
The rest of this section will be devoted to the methodology and conclusions of this second category of empirical privatization research. In the following summary of empirical work from this category of privatization analysis I have attempted to account for the studies and authors that I found to be widely referenced in empirical studies of privatization.
27
Peters & Pierre (2010)28
The World Bank(1995) p. 3729
The World Bank (1995) p. 3730
The World Bank (1995) p. 37One of the most comprehensive studies of post-privatization performance of former SOEs across several fields and nations is the one presented by Megginson et.al in The Financial and Operating Performance of Newly Privatized Firms. Megginson et.al were able to compare the pre- and post privatization performance of 61 former SOEs in 18 countries privatized between 1961 and 1990. The results of this study are relatively uniform and paint a clear picture of the effects on the privatized firms.
31Megginson et.al measured several aspects of the 61 former SOEs such as profitability, efficiency, capital investment spending, output and employment.
Statistical analysis of these categories showed significant increases in nearly all of the above categories.
32The change in the individual categories is listed below.
Profitability: 69.1 percent of the firms showed expanding profit margins and the median increase of the benchmark measure of profitability was 2.49 percent.
Efficiency: The median increase of sales per employee was 12 percent and the median increase of net income per employee was 25.1 percent.
Capital Investment Spending: The median increase of the benchmark measure, capital expenditures divided by sales, was 5.21 percent.
Output: 75.4 percent of the firms experienced increases in sales and the median increase was 24.14 percent.
Employment: 64 percent of firms showed increased number of employees and the median increase was 276 employees. This result was not statistically significant.
In a survey of the theoretical and empirical literature on the differences between public and private ownership, Shirley & Walsh set out to answer a number of questions, one of which regards how SOEs and private firms are affected by market structure and how this may affect privatization outcomes. Performance of firms was measured in terms of operational efficiency defined as “the present value of outputs from a given set of inputs.”
33They find that while both market structure and ownership affect performance, private firms outperform public ones even in competitive markets. In fact, the efficiency increase as a result of privatization is more pronounced in a competitive environment.
34Shirley & Walsh explain that:
31
Megginson et.al (1998) p. 47f32
Megginson et.al (1998) p. 77ff33
Shirley & Walsh (2001) p. 534
Shirley & Walsh (2001) p.10“The empirical literature suggests that while market structure has a positive impact performance, this impact fails to dominate the ownership effect. The argument that market-structure dominance rests on cases in which public and private firms in competitive environments perform equally well, and these cases are rare.”35
In an overview of influential literature on privatizations of SOEs Jordahl presents relatively uniform conclusions regarding the effects of privatization on factors such as efficiency, productivity, profitability, output and debt. The only ambiguity regarding the studies that Jordahl examines stems from the effects of privatization on employment levels in SOEs. The results are especially conclusive when data is available from both before and after a privatization:
“Nearly all studies that assess the same company before and after a privatization find that profitability, efficiency, output, and investments increase while accumulation of debt decreases. How employment at the privatized firms is affected is unclear.”36 (My translation)
A more in depth empirical study of privatizations is Galal et.al’s Welfare Consequences of Selling Public Enterprises. In the study, Galal et.al study privatizations in the United Kingdom, Chile, Malaysia and Mexico in order to answer the question as to whether a country is better or worse off following a privatization of an SOE. In order to discern whether the privatization was beneficial they compare a number of factors relating to the SOEs performance five years’ prior and five years post privatization. They then compare the factual privatization scenario with a counterfactual non-privatization scenario in order to identify any differences that can be traced to the privatization. The factors used to measure the pre- and post-privatization performance of SOEs are: productivity, capital, consumer surplus, two measures of profitability and other returns.
37The empirical research paints quite a uniform picture of privatization resulting in an increase of an SOE’s operational efficiency. Indeed, the World Bank concludes that: “We have presented evidence that SOEs are less efficient than the private sector in competitive markets and, with effective regulation of private firms, in monopoly markets as well.”
3835
Shirley & Walsh (2001) p. 1136
Jordahl, (2008) p. 2337
Galal et.al (1994) p. 4ff38
The World Bank. (1995) p. 50Besides the similarities of the conclusions drawn in the above presented empirical studies there is another glaring similarity, namely that of methodology. The above-described studies all measure the efficiency of the privatization in terms of some sort of comparison between pre- and post-privatization SOE performance. In fact even the measurements used to assess this performance are relatively consistent. For example all of the studies presented above use some form of measurement of productivity, nearly all measure output, profitability and investment, while some also measure employment. The authors whose work is detailed above refer to this work as empirical study of privatization but I believe that a different, more detailed label can shed more light on what has actually been done. It is my view that what is described above, as empirical study of privatization is in fact evaluation; consider the following definition supplied by Vedung. According to Vedung, evaluation is to systematically, post factually, assess the value of an intervention by measuring its effects with the purpose of providing practical guidance going forward.
39By this definition follows that post factually assessing the effects on performance of a privatization in order to assess the viability of privatizations in general is a perfect example of evaluation. Labeling the above presented empirical studies as evaluation raises the specter of evaluation theory and the analytical insight it can provide in the analysis of the above detailed empirical studies.
The realization that the above-presented empirical studies are in fact evaluations gives rise to questions of evaluation theory. Shadish et.al specify five principles and practices that a good theory of evaluation should give comprehensive account of. I would like to focus on the first one which is “What methods to use to produce credible knowledge;”.
40In terms of the above- presented studies; does the measurement of pre- and post-privatization SOE performance produce credible knowledge about the efficiency of privatizations? Rossi et.al refer to this aspect of evaluation theory as construct validity. The concept of construct validity is defined by Rossi et.al as “the extent to which the variables used to measure program constructs convincingly represent the constructs in the program logic model”.
41A program logic model, as mentioned in the citation by Rossi et.al, is a causal model that is designed to plausibly explain how a program is intended to achieve desired outcomes.
42In other words construct validity in this case pertains to the capacity to generalize measurements such as productivity and output to the desired outcomes of the privatization. The concept of construct validity 39
Vedung (2009) p. 19-3340
Donaldson & Lipsey (2006) p. 5941
Rossi et.al (2004)42
McLaughlin & Jordan (1999) p. 2thereby not only stresses the importance of assessing an evaluations operationalization, it also defines the yardstick against which to assess these operationalizations, namely the program logic model, or in other words the goals or purpose of the program itself.
43It is with basis in this principle of evaluation theory that empirical studies of privatization need to be examined. For while the previously detailed parameters do provide significant information about the performance of the SOE I will argue that they are an insufficient base on which to evaluate a privatization and its efficiency. As I will argue in the following section, there is much more to efficiency than simply measurements of performance. The above-mentioned parameters will form the basis for a subsequent evaluation of the efficiency of the privatization of Apoteket AB. This evaluation will provide a baseline that will serve as a comparison to a second evaluation using a different definition of efficiency. The efficiency definition based on the parameters of productivity, output, profitability, investment and employment will henceforth be referred to as operational efficiency. The prefix operational is not used in this context in order to convey any additional meaning but simply to distinguish this definition of efficiency from the one developed in the following section.
An Alternative View of Efficiency
As I have just shown significant weight is given to the measurement of pre- vs. post- privatization efficiency in the above-detailed privatization literature. In this capacity the term efficiency, occasionally associated with the prefix operational, is often used as a synonym for productivity. In this section I will argue that this measure of efficiency provides inadequate information on which to base an evaluation of SOE privatization.
Efficiency is a complicated term with several distinct, as well as changing, meanings and there is even relatively abundant literature devoted entirely to research on the subject. Despite the inherent complexity I will attempt to keep the terminological discussion simple.
Despite the confusion surrounding efficiency, there is some degree of fundamental agreement;
efficiency is the relationship between a cost and achieved change.
44A cost (often in terms of resources and time) is made in order to achieve something (such as manufacturing a car) and 43
Rossi et.al (2004)44
Johansson, (2011) p. 9the efficiency is the relationship between the cost and the car. Since resources aren’t endless, there needs to be some allocation and some form of measurement as to the relative attainment of the invested resources. Unfortunately efficiency, even in a relatively straightforward process such as manufacturing, is rarely as simple as productivity. In other words the number of cars that can be made from a given set of resources is not the only relevant measurement;
productivity only becomes relevant when it is compared to the quality of the output.
45E.g. it doesn’t matter how many cars are made if they aren’t roadworthy. Therefore any serious attempt at measuring efficiency also needs to take into consideration to what extent set goals are met.
The connection between efficiency and goal fulfillment is echoed by Vedung who describes the definition of efficiency as: “The programs capacity to produce predicted outcomes that fulfill the goals of the program (…)”
46From this definition follows that the outcomes that are of interest when measuring efficiency must be produced and predicted by the intervention program and also serve to fulfill the goals of the intervention; or in other words. When analyzing a private sector business the distinction between the use of goals and operational efficiency may not yield any significant differences as one could argue that the goals of such businesses can be more easily defined in terms of economics measurements. Not so for the public sector; the state as owner of an SOE has many concerns and obligations that may not necessarily, for better or worse, coincide with productivity or profit maximization.
47Essentially an evaluation based on operational efficiency runs the risk of becoming an evaluation of the former SOE and not an evaluation of the privatization.
Unfortunately the realization that goals need to be accounted for creates more questions than it answers:
A. Whose goals are important when measuring the success, or lack thereof, of the privatization of an SOE?
B. Are there any formalized goals, and if so what are they?
C. If there are formalized goals, how do we measure them?
45
Oskar Johansson (2011) p. 8-2546
Evert Vedung (2009) p. 3147
Lindgren (2009) p. 48These questions will be addressed in the following section using my chosen case of Apoteket AB as an example.
Whose goals are important when measuring the success, or lack thereof, of the privatization of an SOE?
A: Whose goals to measure performance against, is not necessarily obvious. In fact, when a large SOE such as Apoteket AB is to be privatized a large number of actors are going to have differing goals and visions for the privatization, official or otherwise. In the case of the privatization of Apoteket AB, dozens of organizations, both public and private, were involved directly with the process of shaping the particulars of the privatization.
48Add to that the significant influence of evaluators and investigators and the picture of who in fact is in charge of the privatization process becomes muddled indeed. Ultimately though, the government as the owner formally has the final authority to privatize or not to privatize an SOE. One could argue that government ownership is simply a proxy for ownership by the citizenry but since the citizens are not in the business of formalizing specific goals it would be extremely difficult to achieve any measure of validity when operationalizing these hypothetical goals.
On the one hand one could argue that government does not necessarily have the interests of the citizenry in mind. On the other hand, whether the government actually has its citizens’
interests in mind is not necessarily relevant. What is important is whether the official goals reflect the interests of the citizens and this seems quite likely, as any potential ulterior motives are not likely to be displayed to the general public.
Ultimately it is the government that will take the decision as to whether or not to privatize an SOE and therefore the measuring criterion of privatization success should reasonably have the government’s goals in mind. Simply measuring operational efficiency does not do that accurately. In this case an increase in value of output for a given set of inputs could for example be the result of a price increase, effectively shifting costs from business owners to consumers.
Are there formalized goals and if so what are they?
B: The existence of formalized goals and their explicitness is likely to vary from case to case but it seems unlikely that something as major as the privatization of an SOE would be 48
Prop 2008/09:145 (2009)SOU 2008:4 (2008)
conducted entirely without goals or expectations. Indeed the privatization literature gives plenty of examples of varying goals for privatizations around the world.
49In the case of Apoteket AB there were two clearly formalized goals for the privatization:
1. Increased consumer access 2. Lower prices of pharmaceuticals
Increased consumer access
B1: What is implied by increased consumer access is not self-evident but public documents relating to the privatization effort give a more in-depth description. The factors mentioned that explicitly relate to consumer access to pharmaceuticals are: service hours, number &
proximity of pharmacies and level of service.
The goal to increase service hours is relatively simple and straightforward. It was assumed that a privatization of the pharmaceuticals market and the diversity and competition that would follow would cause pharmacies to have longer service hours than was experienced during the state monopoly as each owner attempts to improve its market share.
50In regards to the number of pharmacies it was assumed that the increased competition and productivity that the privatization created would enable the establishment of additional pharmacies. On the other hand, there was some concern that the sparsely populated nature of certain parts of Sweden would make those areas undesirable to retailers. Prior to the privatization the state monopoly was obliged to provide access to retail pharmaceuticals in all parts of the country and there were concerns that a removal of this obligation would jeopardize access to retail pharmaceuticals in sparsely populated areas.
51Therefore it was decided that one of the conditions for private investors to be allowed to acquire a cluster of pharmacies was that all pharmacies in the cluster were to remain operational for at least three years when terms would be renegotiated.
52The wording used in the formulation of this goal is interesting in the sense that the rural and urban areas are held to different standards. On the one hand the goal is to increase the number 49
Shirley & Walsh (2001)Jordahl (2008)
Megginson et.al (1998)
50
Prop. 2008/09:145 (2008) p. 80f51
Dir. 2006:§136 (2006) p. 552
WP/PM 2011:49 (2011) p. 18of pharmaceuticals retailers but on the other hand the goal is also to “safeguard” the access to pharmaceuticals in sparsely populated areas.
53The nature of the word safeguard (Swedish:
trygga) implies protecting or preserving. This implies that the deciding factor for rural areas may not be the number of pharmacies as much as guaranteeing that there is at least one pharmacy that can be considered reasonably accessible.
As opposed to the other two factors of consumer access, level of service is quite ambiguous.
Upon closer examination of public documents it becomes clear that level of service pertains primarily to the ability of the pharmaceuticals retailers to live up to the obligation of supply that applied to the state monopoly. Apoteket AB was required to supply all nationally and European Economic Area (EEA) approved medication to consumers within 24 hours.
54There was special concern that this capacity may be threatened primarily in rural populated areas.
55Lower prices of pharmaceuticals
B2: There are two distinct parts of the goal of lower prices on pharmaceuticals: the cost to consumers and the cost to government.
Part of the basis for the goal of lowered prices of pharmaceuticals is the fact that prior to the privatization the prices paid to manufacturers by Apoteket AB for pharmaceuticals (Manufacturer to Pharmacy Price (MPP)) with active patents was the second highest in the OECD.
56This was seen as a significant problem as active patent pharmaceuticals are by far the largest cost driver, accounting for approximately 70 percent of the value of all sales.
57It was assumed that private businesses would be more effective at negotiating the price of these pharmaceuticals and that this would create the possibility for reduced pharmacy-to-consumer prices (PCP) through lowered MPP.
58While the goal to reduce the prices of pharmaceuticals does not specify at what stage of the process from manufacturing to retail it aims to reduce prices, the fact that the supposed beneficiaries are both consumers and government leads me to believe that PCP is the price that is to be lowered. But while reducing the price is all well and good, a reduction of price is not the actual purpose but the means to an end. In clearer 53
Dir. 2006:§136 (2006) Page 154
Moïse & Docteur (2007) p. 4355
SOU 2008:4 (2008) p. 477f56
SOU 2008:4 (2008) p. 3357
SOU 2008:4 (2008) p. 33f58
SOU 2008:4 (2008) p. 37terms this means that the goal is not to in fact necessarily to reduce the prices paid but to reduce the aggregate cost. Any reduction in price that does not reduce the actual cost is entirely meaningless and could for example be the result of deflation or other circumstances.
But that is not to say that price information is entirely useless as it may prove useful as part of an analysis of cost fluctuation.
The costs for consumers and government for retail pharmaceuticals are connected through a system of cost sharing. Consumers pay for all costs of prescription medication up to a cap after which government pays for all subsequent costs. Non-prescription medication on the other hand is generally not subsidized and the consumer pays in full.
59This system has historically resulted in the consumer paying for approximately 30 percent of the total costs of retail pharmaceuticals with the state and local governments paying the remainder of the costs.
60As the cost-sharing scheme does not apply equally to all types of medication this creates a situation where primarily consumers enjoy any cost reduction in non-prescription medication. Any cost reductions in prescription medication will primarily benefit government as consumers already above the cap make the majority of the purchases of prescription medication from retailers.
61How do we measure the fulfillment of the formalized goals?
C: How to measure the goals of the privatization is important to determine in order to reliably assess the efficiency of the privatization. In order to create simplicity and structure my chosen assessment parameters will culminate in a number of hypotheses that will aid the process of analyzing the empirical results.
Increased consumer access
C1: The goal of increased consumer access has three parts: service hours, number and proximity of pharmacies, and level of service. The varying nature of these factors does not lend itself well to assessment as a whole, rather I believe each factor needs to be assessed individually and the results compiled in to a general evaluation of goal fulfillment.
Measuring potential changes in service hours of pharmacies across the country pre- as well as post-privatization should satisfactorily assess the goal fulfillment of increased service hours of 59
Moïse & Docteur (2007) p. 1560
SOU 2008:4 (2008) p. 32f61
SOU 2008:4 (2008) p. 111pharmacies. Detailing any potential change in service hours for all pharmacies individually is obviously quite unwieldy. With this in mind an average change value, if any, seems adequate to assess goal fulfillment. Data detailing the average service hours of retail pharmacies pre- and post-privatization was obtained from the Swedish Pharmacy Association. For the goal of increased service hours to be fulfilled it is my assessment that the average service hours should have increased as a result of the privatization. This leads us to the first hypothesis:
#1 H0: The average service hours of pharmacies have not increased post privatization.
Number and proximity of pharmacies has two parts. The general number of pharmacies can be assessed with relative ease; it is simply a matter of collecting numerical data from before and after the privatization and comparing the two. But in order to attribute an increase in the number of pharmacies to the privatization we also need to put any change in the number of pharmacies in relation to any trend in the number of pharmacies that might have existed prior to the privatization. For example, if the number of pharmacies in Sweden was growing by an average of 50 pharmacies per year prior to the privatization, a continued increase of 50 pharmacies per year post privatization would not be attributable to the privatization. For the goal of an increased number of pharmacies to be fulfilled it is my assessment that the aggregate number of pharmacies must have increased post privatization when compared to the projected change of the years leading up to the privatization. Data detailing the number of retail pharmacies for the various years was obtained from two separate sources. Apoteket AB supplied the pre-privatization data while the post-privatization data was obtained from The Medical Products Agency. This leads us to our second hypothesis:
2# H0: The aggregate number of pharmacies has not increased post-privatization compared to the average change in the number of pharmacies of five years prior to privatization.
The question of proximity on the other hand is a different matter. In order to measure the goal fulfillment of safeguarding the access to pharmaceuticals in rural areas the data needs to reveal the number of pharmacies located in these areas before and after privatization.
Unfortunately the number of pharmacies in rural areas alone does not necessarily accurately
portray the average citizens access to pharmacies, as they may not be spread evenly across
these areas. To account for this problem I will be using a measure the number of people that
have to travel more than a certain amount of time by car in order to access a pharmacy. This
measurement is based on an area type model created by the Swedish Rural Agency (Swedish:
Glesbyggdsverket). This model has been used in numerous studies in order to describe changes in access to various services and markets.
62The data detailing travel time to the nearest pharmacy based on the area type indexation model was obtained from The Swedish Agency for Growth Policy Analysis. In addition to this there is also the matter of measuring the relative vulnerability of this access. One way of assessing the vulnerability of consumer access to pharmaceuticals would be to measure what effect the closing of the nearest pharmacy would have on the required travel time. In other words, a small difference between the distance to the nearest pharmacy and the second nearest pharmacy would indicate low vulnerability and vice versa. Again, data for all these factors needs to be acquired from before and after privatization and trended where necessary. The data detailing the change in travel time that would result from a hypothetical closing of the nearest pharmacy was also supplied by The Swedish Agency for Growth Policy Analysis. For the goal of increased pharmacy access in rural areas to be fulfilled it is my assessment that the number of pharmacies in rural areas needs to have increased post privatization when compared to the projected change of the years leading up to the privatization. It is also my assessment that the average travel time to the nearest pharmacy should have decreased as a result of the privatization and that the travel time to the second nearest pharmacy should have decreased as well. This leads us to our third, fourth and fifth hypotheses:
3# H0: The number of pharmacies located in rural areas has not increased post privatization compared to the average change in the number of pharmacies of five years prior to
privatization.
4# H0: The average travel time required to reach the nearest pharmacy has not decreased post privatization compared to the average change in the number of pharmacies of five years prior to privatization.
5# H0: The average travel time required to reach the second nearest pharmacy has not
decreased post privatization compared to the average change in the number of pharmacies for five years prior to privatization.
62
WP/PM 2011:49 p. 12fLevel of service is a quite intangible concept that needs to be focused into a defined
measurement that lends itself to assessment more easily. The goal of increased level of service is not explained extensively in public documents but seems to pertain primarily to the
capacity of retailers to deliver all nationally and EEA approved medication to consumers within 24 hours. For obvious reasons detailing the success rate of medication availability within 24 hours of every pharmacy individually is impractical. Therefore it is my assessment that an average success rate of availability of sought medication within 24 hours post
privatization compared to the pre privatization equivalent would be the best way to measure the fulfillment of this goal. The data detailing the success rate of delivery of sought
prescription pharmaceuticals at the first visit pre-privatization was supplied by the Dental and Pharmaceuticals Benefits Agency and post-privatization data by the Swedish Pharmacy Association. This leads us to the sixth hypothesis:
#6 H0: The average success rate of delivery of sought prescription medication within 24 hours has not improved post privatization.
Lower prices of pharmaceuticals
C2: There are two distinct parts of the goal for lower cost of pharmaceuticals: the cost to consumers and the cost to government. In order to gather as much information as possible from which to assess the fulfillment of the goal of lower cost of pharmaceuticals it is prudent to examine both prescription and non-prescription medication. In order to assess the change in the cost of pharmaceuticals to consumers and government the pre-privatization cost trend must also be taken in to consideration.
Unfortunately there are some problems trending the costs of prescription medication. Because
of the fact that a significant amount of the cost of pharmaceuticals can be derived from
pharmaceuticals with active patents, the expiration and introduction of patents on
pharmaceuticals with widespread use cause large fluctuations in the aggregate cost of
pharmaceuticals. The cost-sharing scheme causes this effect to have major impact on the costs
for government but a much lesser impact on the cost for consumers as consumers of
expensive patented pharmaceuticals will quickly reach the consumer payment cap. The nature
of this effect means it does not lend itself well to trending as it occasionally causes large
fluctuations from one year to the next. Fortunately, while this does lend a certain degree of
uncertainty to the cost development results, patent introduction or expiration dates are not
complete unknowns and should be able to be accounted for to some degree when the empirical data is analyzed.
In order to fulfill the goal of reduced pharmaceuticals costs it is my assessment that the costs of non-prescription and/or prescription medication should show reduction compared to the trend of the past few years. The data pertaining to cost development of pharmaceuticals pre- privatization was obtained from Official Reports of the Swedish Government while the post- privatization data for aggregate pharmaceuticals cost was obtained from Pharmaceutical Service AB while non-prescription cost data was obtained from The National Board of Health and Welfare. Pharmaceuticals consumption data was supplied by Pharmaceutical Service AB.
This leads us to the seventh and eighth hypotheses:
7# H0: The costs of prescription pharmaceuticals have not decreased post privatization compared to the average change of the years previous to the privatization.
8# H0: The costs of non-prescription pharmaceuticals have not decreased post privatization compared to the average change of the years previous to the privatization.
What, besides goal fulfillment, is needed in order to assess efficiency?
An SOE, or any other business for that mater, does not operate in a vacuum, there are also societal costs associated with privatization of SOEs. While the degree of interference in private business may vary, nearly every political system is likely to have some form of government supervision of private business. Regulation and supervision are likely to increase with a decrease in direct government involvement in favor of private alternatives.
63This supervision, regardless of its scope, is not free and if it is likely to increase. This cost increase should be accounted for when measuring the general success of the privatization.
Methodology – Measuring the Effects of Privatization, Dual Approaches
63
Jessop (2010) p. 64ffThe hypotheses pertaining to increased consumer access are, as demonstrated in the section above, the following:
#1 H0: The average service hours of pharmacies have not increased post privatization.
#2 H0: The aggregate number of pharmacies has not increased post privatization compared to the average change in the number of pharmacies during three years prior to privatization.
#3 H0: The number of pharmacies located in rural areas has not increased post privatization compared to the average change in the number of pharmacies during three years prior to privatization.
#4 H0: The average travel time required to reach the nearest pharmacy has not decreased post privatization compared to the average change in time required during three years prior to privatization.
#5 H0: The average travel time required to reach the second nearest pharmacy has not decreased post privatization compared to the average change in travel time during the three years prior to privatization.
#6 H0: The average success rate of delivery of sought medication within 24 hours has not improved post privatization compared to pre-privatization.
And the hypotheses pertaining to the goal of lowered prices of pharmaceuticals are:
#7 H0: The costs of prescription pharmaceuticals have not decreased post privatization compared to the average change in costs during years previous to the privatization.
#8 H0: The costs of non-prescription pharmaceuticals have not decreased post privatization compared to the average change in costs during years previous to the privatization.
Input
While the above hypotheses are designed to measure goal fulfillment there is another
parameter that needs to be taken into consideration in order to measure the efficiency of the
privatization and that parameter is input. As per the efficiency definition in the previous section, goal fulfillment needs to be contrasted to the size of the cost that was paid in order to achieve the stated goals. The primary sacrifice that was made to achieve the privatization goals was in terms of increased workload for state agencies and a resulting increase in costs.
There are three state agencies which were deemed to be directly affected by the privatization of the Swedish pharmaceuticals monopoly (Swedish translations in parentheses): The Medical Products Agency (Läkemedelsverket), The Dental and Pharmaceutical Benefits Agency (Tandvårds- och läkemedelsförmånsverket) and The National Board of Health and Welfare (Socialstyrelsen).
64The Medical Products Agency was deemed to have to bear the brunt of the workload increase resulting from the privatization. Among other new tasks, the new process of issuing licenses to prospective retailers was deemed a significant burden:
”With the suggested system for licensing application, notification of pharmacy establishment and notification of changed circumstances, The Medical Products Agency will need to administer and assess these applications and notifications. This will be a completely new task for The Medical Products Agency which, depending on when and to what extent new retailers seek to establish themselves on the market, will require considerable resources both in terms of manpower and restructuring.”65 (My translation)
In addition to the above-mentioned costs the privatization was deemed to require significant increases in supervision of pharmacies as well as of pharmaceuticals imports. Many of the tasks that the agency performed prior to the privatization and continues to perform post privatization also needed to be restructured. All of The Medical Products Agency’s cost increases will however be financed through fees imposed upon the private retailers and will therefore not burden government spending.
66The Dental and Pharmaceutical Benefits Agency on the other hand has its day-to-day operation financed by the state government and it was also projected to receive an increased workload.
67The agency was primarily tasked with monitoring and to some extent evaluating 64
SOU 2008:4 p. 581ff65
SOU 2008:4 p. 58266
SOU 2008:4 p. 58267
SOU 2008:4 p. 582fcertain aspects of the systems for exchanges of pharmaceuticals with generic competition and imports of specialty pharmaceuticals. The primary task of the agency, which was to organize the prescription pharmaceuticals price cap scheme, will remain unchanged. This will keep the bulk of the costs for the agency unchanged. Apart from certain one-off costs, which will be discussed below, the agency was to receive an increase to its annual budget of SEK 5-10 million (USD Approx. 1.1 million).
68The National Board of Health and Welfare has the supervision of pharmacies as its primary responsibility in relation to the pharmaceuticals market. The expected increase in the workload was explained as follows:
“The pharmaceuticals market is not seen by the agency as a large patient safety risk but the large quantities of pharmaceuticals that are handled still require certain resources to be allocated in order to ensure safety. The supervision is more dependent on the organizational form of, rather than on the sheer number of, pharmacies. Apoteket AB had a well functioning safety system functioning identically across the country which allowed supervision to be systematized.”69
