An Empirical Investigation of Factors Affecting the Standardization of Service Pricing:

An Empirical Investigation of Factors Affecting the Standardization of Service Pricing:

___________________________________________________________________________________

A Case Study of Flexlink

Louise Lorentzon & Oscar Johnsson

2013-05-23

Abstract

Over the last decades, the service sector has increased as a percentage of the economies. This paper analyzes whether service prices can be internationally standardized based on the Purchasing Power Parity (PPP) theory. The findings confirm that the services’ unique features intangibility, inseparability, heterogeneity and perishability (IIHP) complicate standardization, for an international price comparison to be possible. From our models, there is no statistical proof that exchange rate, as the sole price adjustor in PPP, has an effect on pricing. We conclude that there are macro factors in the models, not covered by the PPP theory that acts against an implementation of standardized prices. From the study on Flexlink, a company that assembles and installs transporter systems, it was verified that the sales units regularly use price discrimination, as a result of the markets’ high price elasticity, an adaption to the local markets are necessary in order to be competitive.

Keywords: Pricing, Services, PPP, IIHP, FTU, International trade, Wage, Cost of Living

Spring 2013 Supervisor: Martin Holmén Bachelor Thesis in Corporate Finance (15 hp) The Department of Economics at the School of Business, Economics and Law Oscar Johnsson 19900606-0231, Louise Lorentzon 19890518-0140

Acknowledgements

We would like to thank our supervisor Martin Holmén for sharing his knowledge, and for helping us with the issues that arose during the writing of this thesis. We would like to thank Carina Karlsson, Magnus Andersson and Kristina Wall Jungbjer for assisting us within each of their area of professional expertise.

We would also like to give thanks to all the Finance Managers who participated in the interviews.

And finally, we would like to thank Hilda, who entertained and inspired us during our long working days.

________________________ __________________________

Oscar Johnsson Louise Lorentzon

ACKNOWLEDGEMENTS 2

1. INTRODUCTION AND PROBLEM STATEMENT 5

1.1. Background 5

1.2 Research Questions 7

1.3 Methodology and Delimitations 7

1.3.1 Literature Review 8

1.3.2 Survey 8

1.3.3 Regressions 8

Fig.1 Essay Process: The Figure Shows the Main Structure of the Thesis. 9

2. FLEXLINK’S HISTORY 9

3. DEFINITION AND FRAMEWORK 10

3.1 Definition - Service 11

3.1.1 Intangibility 12

3.1.2 Heterogeneity 13

3.1.3 Inseparability 14

3.1.4 Perishability 14

3.2 FTU Framework 15

3.2.1 Facilities 16

3.2.2 Transformation 16

3.2.3 Usage 17

4. THEORY 18

4.1 Purchasing Power Parity (PPP) 18

4.1.1 Definition 18

4.1.2 Law of One Price 19

4.1.3 The Development of the PPP 19

4.1.4 Critiques of the Applied PPP 20

5. METHODOLOGY 22

5.1 Macro Variables on Pricing 22

5.1.1 Dependent Variables 22

5.2.3 Independent Variables 23

5.3 The Regression Model 27

5.4 Hypothesis 28

6. EMPIRICAL RESULTS 29

6.1 Correlation Results 29

6.2 Regression Results 31

6.3 Analysis of Regressions 34

6.3.1 Cost of Living Models 34

6.2.2 Cost of Engineer Models 35

6.2.3 Cost of Fitters Models 36

6.3 Survey Results 37

6.4 Analysis Survey 39

7. DISCUSSION 40

8. CONCLUSIONS 42

9. LIMITATIONS AND SUGGESTIONS 44

BIBLIOGRAPHY 46

Literature and Articles 46

Websites 49

Interviews and Observations 51

APPENDIX 52

Definitions

AGRI_% Percentage of the population working in the agriculture sector UNEMP_% Percentage of the population unemployed

BMI_PPP Big Mac index – Purchasing Power Parity adjusted GDP/CAP The GDP/ Capita in Dollar

GDP_DEF The GDP Deflator

COL_IN Cost of Living index, with USA as base country EX_RATE Exchange rate, with SEK as base currency

POVU_2_% Percentage of the population earning less than 2 dollars/day SERV_% Service sectors percentage part of the GDP

AV_WAGE Average wage/hour

PPP Purchasing Power Parity

IIHP Intangibility, inseparability, heterogeneity and perishability FTU Facilities, transformations and usage

OECD Organization for Economic Co-operation and Development ENG_COST Average Engineer cost/hour at Flexlink

FIT_COST Average Fitter cost/hour at Flexlink

GDP Gross Domestic Product

CPI Consumer Price Index

1. INTRODUCTION AND PROBLEM STATEMENT

1.1. Background

Organizations that pursue business globally, are participating in several markets. If a customer has affiliates in more than one country that require services, prices should be internationally standardized to accommodate the trade. International firms want to charge the same price to the same customer in different countries; otherwise, the same customer will have different costs for different countries that will incur problems.

Based on these conditions, standardized prices ought to be the norm, but there are several factors that complicate the standardization of service performance and price.

Service pricing is becoming an increasingly important topic for multinational corporations, as there have been a shift from products to services in developed countries. Cassel (1918) presented the Purchasing Power Parity (PPP) theory, based on the principle of equal purchasing power in all countries after adjusting for exchange rate. The PPP theory by Cassel (1918), as well as Taylor and Taylor (2004) and Rogoff’s (1996) take on the theory, provide the framework for this thesis concerning international standardization of service pricing. A company’s main objective is to maximize profit by using the best strategy. Based on the theory by McCarthy (1960) there are 4 P’s; product, price, place and promotion, all of which are costs except for price. To set the right price is of utmost importance since a too high price might make the firm non-competitive, while a too low price can create losses.

Furthermore, existing theories and literature mainly consider pricing of products, and there is an absence of thorough studies on services.

This paper will discuss the subject from an economic and management view. Starting from a macro level with empirical results, there will be three cases examined; a general for the world, for fitters and for engineers. The service professions differ in terms of difficulty; fitters assemble the systems, while engineers design and act as project leaders. The macro factors applied are recognized as good indicators for pricing. At a micro level, a case study of Flexlink, a company that assembles and installs transporter systems.

Current research and literature is vague and ambiguous regarding the difference of goods and services, and therefore there is an attempt to define a service and the specific characteristics. The aim of this thesis is to examine whether service prices can be internationally standardized. According to the PPP theory there should be global price parity, and if violated it provides evidence against internationally standardized service prices.

The case study will be conducted on Flexlink; a global company, headquartered in Gothenburg, that assembles and installs transporter systems. Their revenue is based on selling projects that include the product and service. The regression analysis will be based on Flexlink’s cost of services since they have reported differences among their 19 sales units, and in-depth interviews will be conducted. The case study will favourably confirm the findings on a global level.

Pricing theory is a complex subject due to the many aspects involved when setting the price, but by a multifaceted approach the aim is to give a comprehensive assessment.

The different perspectives presented in this paper are an attempt to contribute to price theory, with a focus on services; from a macro to a micro point of view.

1.2 Research Questions

(i) Can service prices be standardized based on economic theories or should an adaption be made to the local market?

(i.i) Do macro factors affect the pricing of services such that standardized pricing is not feasible?

(ii) Can the Flexlink findings strengthen the Purchasing Power Parity for service pricing?

1.3 Methodology and Delimitations The thesis has two approaches:

Macro: Examined through literature and empirically.

Micro: Case study; survey results to strengthen the findings.

The assessment of the perspectives will be carried out by using a survey, regression and an analysis of the literature.

1.3.1 Literature Review

In order to answer the research questions, the characteristics of a service need to be separated from the goods. First the service is discussed with the traditional IIHP (Intangibility, heterogeneity, perishability and inseparability), a model promoted by Lovelock (1999), Bowen and Schneider (1988), Gummesson (2004), Berry (1980), Beaven and Scotti (1990). From critiques of the IIHP, the more recent FTU Framework is introduced with a process based approach (Moeller, 2008; Fließ and Kleinaltenkamp, 2004; Vargo, 2008). However, both models can be seen as simplifications, but are necessary in order to identify the service characteristics.

The thesis main theory is the Purchasing Power Parity (PPP) that supports the standardization of service pricing, if price parity exists (Cassel, 1922; Taylor and Taylor, 2004). The theory is a valuable tool for explaining pricing, but has been criticized to not hold in a globalized world.

1.3.2 Survey

The survey is based on interviews with the Finance Managers in 7 out of 19 countries where Flexlink have sales units. Due to the short time window for the survey interviews, all countries could not participate. The questions are divided into two parts;

internal customers are defined as in-house trade between Flexlink’s sales units, and all other trade is with external customers, non-Flexlink. The dividing of questions into two parts (Appendix), was done after input from the Flexlink supervisors to make it easier for the sales units to give correct answers. There are a total of 31 questions, both multiple-choice and full-answer. The Finance Managers provide an updated and reality based view of service pricing and the questions cover topics from macro, micro to accounting (Appendix).

1.3.3 Regressions

In the regression tests, panel data is used. The data is collected for the years 2001- 2011. In the world regressions, the dependent variable is Cost of Living, and in the Flexlink regressions, the dependent variables are Cost of Engineers and Cost of Fitters.

The independent variables are: Big Mac Index, Agriculture, Poverty limit $2, unemployment, GDP/Capita, GDP Deflator, Exchange rate, average wage/hour and percentage employed in the service sector. Below are the countries where the nineteen sales units are located:

Europe Belgium France* Germany Hungary* Italy

Poland* Spain* Sweden* United Kingdom Russian

Asia and the Pacific Australia* China India Indonesia Malaysia* Singapore North and South

America Brazil Canada United States

*) The countries chosen for personal interviews

Fig.1 Essay Process: The Figure Shows the Main Structure of the Thesis.

The process of the essay starts from an Informatory point of view, where the authors’ describe and investigate the problems, limitations and main approaches. Thereafter, the essay’s perspective is an Exploratory view and focus is on relevant literature. After that, an Empirical aspect is applied to analyse the regressions and survey results in order to answer the research questions. Finally, the findings are analysed and evaluated in order to confirm the results.

Informatory

Informatory Exploratory Exploratory Empirically Empirically Analysis Analysis Confirmatory Confirmatory

2. Flexlink’s History

Timeline

Flexlink is a manufacturing based company of conveyor system; the procedures are loading, processing, assembly and packaging. In 1980, Flexlink was founded as a spin-off company from the roller ball bearing maker SKF (Svenska Kullagerfabriken AB). The international success came immediately, and the Flexlink conveyor systems quickly got recognition in a wide range of industries (Flexlink 2013). At present, Flexlink is one of the main players in the segment material handling systems, and has established Sales Units in 19 countries and partner companies in over 50 countries (Flexlink information PPT 2013). The historical timeline is based on Flexlink’s documents (2013) and Flexlink’s informational PowerPoint (2013).

1980

1980 Flexlink was founded in Gothenburg, as an efficiency project within SKF

1982-87 Development of the organization and product solutions

1988–03 Establishment of Sales Units in:

Japan, USA, Germany, Singapore, Brazil, France, Italy, UK, the Netherlands, Australia, Czech Republic, China, Poland, Hungary, Spain, Malaysia, Thailand, Finland. The first information/co- ordination system was introduced

1997 Flexlink was aquired by EQT and became an independent company for the first time

2005 Acquired by ABN Amro Capital, now AAC Capita

2006 Acquisition of Tops Conveyors, a company based in Canada

2007 Opening of sales companies in India and Indonesia

2010 Acquistions of the

Automation Division from Schüco,

Germany, and Italian e-cube. 2011 January 11^TH FlexLink was acquired by COESIA GROUP, and became a member of the innovation based industrial group from Bologna, Italy.

2012 Flexlink is one of the largest specialized conveyor distributers in the world with revenues of 1,6 billons SEK

3. DEFINITION AND FRAMEWORK

3.1 Definition - Service

During the past 20-30 years, there has been a dramatic increase in the service sector.

According to Edvardsson et al. (2000), different services in the private as well as in the public sector account for about 70-75 % of the gross national product in most OECD-countries. He also argues that services are gaining in importance and are gradually becoming an essential part in all types of commodity manufacturing firms;

many firms even define themselves as service oriented rather than engineering oriented. Nowadays, it is a common understanding among most service researchers that in a world characterized by turbulence (Mintzberg, 1993), services are becoming an outstanding tool both to differentiate the business (Oliva & Kallenberg, 2003) and to develop sustainable competitive advantage (Barney, 1991).

It has been debated, for instance by Matanovich (2003), that there should be no difference between pricing of goods and services. He emphasizes the importance of focusing on how much value that is offered to the buyer rather than focusing on if it is a good or a service. However, there is more evidence proving that key differences exist between the pricing of goods and services, and this will be our view for this thesis.

Lovelock (1999) argued that the rather intricate nature of the service sector has created a complex problem regarding the definition of a service. Not only have services gained importance in the developed world, but what is perceived as a service has also changed due to technology (Rust, 2004). The out-dated but still general view is that services are based on personal contact (Bowen, 2000), but many services are nowadays net-based e.g. personal banking, lectures, shopping. In order for the term

“service” to be defined correctly, new traits and sectors of application must be added to make it current (Moeller, 2010).

Lovelock(1999) argues that it is the performance/delivery of the service that makes it so hard to grasp, and the fact that services are quite abstract. To most people, he argues that to understand the manufacturing businesses is much simpler. You can easily understand the process of assembly, transformation or creation of goods, by

using physical input to get physical output in for example a factory. To understand or define what type of value, or how much a service really matters is much harder according to Lovelock (1999).

From findings in the work of other researchers there is a wide range of methods in order to distinguish goods from services. Given the examined research material, two approaches will be discussed. Firstly, the characteristics of a service are condensed into four main groups, IHIP; Intangibility, heterogeneity, perishability and inseparability (Lovelock, 1999), (Hansen & Mowen, 2006), (Barney 1991), (Flipo 1988) followed by a more modern approach, the FTU Framework (Moeller, 2008).

Fig. 3 - IHIP Characteristics (Explained below)

3.1.1 Intangibility

The intangibility aspect is always connected to services and has for long been considered as the most fundamental characteristic (Bowen and Schneider, 1988). The intangible side described; cited Berry (1980, p.24) “A good is an object, a device, a thing; a service is a deed, a performance, an effort”. The intangible side of a service is subject to its immateriality and the very problem when it comes to describing what a service stands for, and the valuation of the performance. This problem is not new, back in the 16^th century; Smith (1776) felt that services were non-value adding and only goods that could be included in trade would create wealth for the country. Say (1836) considered Smith to be wrong and argued that even though services are not

Intangibility

Intangibility HetrogeneityHetrogeneity

Perishability

Perishability InseparabilityInseparability

material, their usefulness and surrounding activities create value. The difficulty of selling a service is that what is being sold is a contract of an immaterial performance (Bowen and Schneider, 1988). What the buyer actually receives and the quality of the service is vague which causes uncertainty regarding the value (Shostack, 1977).

There is criticism of the intangible aspect of a service. In order to perform a service, tangible goods are normally required (Shostack, 1977). An obvious situation is a barber performing a haircut, perceived as an intangible service. For the service to be executed there are a need for a studio, scissors, hairdryer which are tangible. A service performed on a person who is tangible (haircut), will result in a change due to the service and is therefore tangible (Moeller, 2010). Much focus has been placed on the transformation process of resources given by the customers' input, which also can be interpreted as the services’ main purpose (Hill, 1977). For most goods input is not given by the customers for transformation and only the finished product is sold.

However, there are grey areas in terms of customized goods, when the customer asks the seller to perform a service that results in a product (Moeller, 2010).

3.1.2 Heterogeneity

One of the difficulties regarding services is for the outcome to be the same, and thus to standardize a service is hard (Edgett an Parkinson, 1993). Different factors that are considered to be making the service heterogeneous have been investigated with separate viewpoints. Beaven and Scotti (1990) suggested that it was the outcome that differed and thus the services’ results cannot be considered homogeneous. According to Lovelock and Gummesson (2004) it also depends on how the service is performed and that this creates a heterogeneous situation. The same service’s quality can vary a lot depending on who performs it and the daily status of that person. It does not only depend on the service provider for the service to be heterogenic, the customer expectation and participation plays an important role in the service outcome (Palmer and Cole, 1995).

Not everyone agrees that services cannot be homogeneous, it depends on the nature of the service and some may therefore be standardized (Lovelock and Gummesson, 2004). According to Moeller (2010), some services can be standardized by using the same process, while others such as customized services cannot be made by following

the same model. There are discussions concerning whether it is the service that actually varies, but according to Vargo and Lusch (2004) the variations are due to the human error, which can be reduced but not avoided. The technology has become more advanced over the past centuries, and services previously handled over the counter have become computer-based, e.g. banking and shopping (Furrer, 2003). The result of this is that more services can be standardized and variations will decrease according to Lovelock (1983). The input to the service provided by the customer will vary, and therefore the amount of time and the execution might differ but this is not due to the service offered (Lovelock, 1983, p.16).

3.1.3 Inseparability

Inseparability comes from the fact that the production and consumption of a service takes place at the same time, making it impossible to divide the two (Say, 1836).

Berry (1980, p.259) argues that this is due to the fact that the customer and the person performing the service cannot be separated and both must be in the same place at the same time. From this follows that the relationship between the provider and customer plays an important role for the perceived service experience and overall impression (Bitner, 1990). Other differences between services and goods have also been described, such as the order of the sales process; the service is sold in terms of an agreement, and then the service is produced and consumed at the same time.

Lovelock and Gummesson (2004, p.29) found that there are many services provided in today's society where the customer is not present at the process of the service. They gave examples of transport companies carrying goods, but when the customer receives the packet the service is already completed. However, Lovelock (1983) argues that services are sometimes characterized by inseparability but this depends on the service being performed. He chooses to split up the services into two groups, the physical body and physical belongings. The former must be present at the execution of the service but regarding the latter, the customer does not need to be present.

3.1.4 Perishability

This is one of the things that distinguish services from goods; they disappear while they are consumed. It is one of the main opinions regarding why services are difficult to analyse, the fact that the service cannot be kept or stored (Beaven and Scotti, 1990;

Kotler, 1994). Smith (1776, p. 351) was straight to the point on perishability “the

labour of the mental servant, on the contrary, does not fix or realize itself in any particular subject or vendible commodity. His services perish in the very instant of their performance”.

The criticism of perishability is directed against the fact that services are available in computer systems, knowledge and people (Gummesson, 2004, p.124). Services are also preserved in the form of receipts showing the performed action (Edvardsson et al., 2005). Lovelock (2000) argues that services are different and some has a time- limit while others continue to give value to the customer. The perishable aspect is not unique for services; many goods disappear after consumption e.g. food. From some services, value remains after consumption and therefore what is left behind is a proof that services do not always completely perish. This is very obvious regarding education, the service performed by the professor of teaching the students, will remain as knowledge (Hill, 1977). It is hard to adapt services to the economic cycles since services cannot be stored and are entirely based on customers’ demand when using a facility (Ng et al., 1999).

3.2 FTU Framework

It is not an optimal approach to describe what differentiates a service from a product by using the four IHIP characteristics (Lovelock and Gummesson, 2004, p.32). The reason is that society has changed and today the research focuses on other kinds of services, and not only the personal based services that were more occurring in the past (Bowen, 2000). The IHIP characteristics are seen as valuable parameters to describe a service (Edvardsson et al, 2005), but there is a need for adaption to better fit today's society. The problem according to Moeller (2010) is that the implementation of IHIP characteristics is vague, and that the different characteristics need a new interpretation of what part and stage of the service is described. To do this, Moeller (2008) argues that services should be divided into different stages and thus get the four IHIP characteristics to work better as helping tools to define a service. One way to do this is to use the FTU framework that describes a service in three process parts facilities, transformation and usage.

Fig. 4 FTU Framework; IHIP Characteristics

3.2.1 Facilities

Facilities enable the service performance and are the value-required input in the processes (Moeller, 2010). Facilities are usually a workplace where the service can be performed, machines and equipment, and staff with knowledge (Shostack, 1992).

Edvardsson et al. (2000) phrased it that these are essentials for a service to take place.

Under these conditions the facilities consist of both tangible and intangible assets and it is clear that a service cannot completely be associated with intangibility. Facilities are only the input for the service to be performed, without customers, the assets remain unused (Fließ and Kleinaltenkamp, 2004) as opposed to goods that can be produced and stored.

According to Moeller (2010) perishability is a characteristic of facilities since the competence (personnel and equipment) provided, must be supported by a customer’s input of resources, in order to be activated and not perish as seen from the provider’s point of view. Heterogeneity is also a part of the facilities regarding what the customers provide, their input will vary and so the prerequisites are not homogenous (Moeller, 2010).

3.2.2 Transformation

The service that is sold is a promise of a change to fix or improve (Hill, 1977).

Depending on which resources are used, these can be divided into direct and indirect service provisions (Moeller, 2008). First, indirect service provision is when the provider arrange for the input to the service in the case of goods. Regarding the direct

Perishability Heterogeneity

Facilities

Intangibility Inseparability

Transformation

Value-creation

Usage

service provision is when it is the customer who provides input to the service, and what we traditionally call a service. In the latter case the limitation of the provider is that the customer is owner to part of the resources necessary for the service that cannot be accessed otherwise (Hill, 1977). The resource can be the client's body e.g.

surgery or an asset like car reparation or knowledge-based banking services etc. (Fließ and Kleinaltenkamp, 2004). The customers contribute their resources because of the need for expertise, along with the providers assets a transformation can be achieved (Moeller, 2010). Vargo (2008) argues that there are two ways in which the provider interacts with the customer; coproduction is when the two work together to create the transformation, and co-creation of value is when synergies create added value to both parties.

In the transformation part what the provider sells to the customer is an agreement of a service and not a tangible product. The intangibility aspect is therefore a part of the transformation stage (Moeller, 2010). In order for the service to be performed in the transformation stage, the customer’s resources needs to be present and are therefore inseparable from this stage (Moeller, 2010).

3.2.3 Usage

The usage of the transformed service is the last step in the FTU framework (Fließ and Kleinaltenkamp, 2004). In order to create value the customer needs to use the service to benefit from the improvement (Moeller, 2010). The provided service that is now used can either be standardized based on a template from a catalogue or customized to fit the customer's unique needs (Fließ and Kleinaltenkamp, 2004).

4. THEORY

4.1 Purchasing Power Parity (PPP)

The Purchasing Power Parity (hereafter PPP) has been chosen as main theory for this thesis, and all discussions will be based on either PPP or the macro factors examined in the next part of the paper. The PPP theory’s assumptions are valuable in order to connect the literature and previous findings, and answer the first research question

“Can service prices be standardized based on economic theories or should an adaption be made to the local market?”.

This part of the thesis will examine and describe the PPP theory, evaluate other researchers’ findings, and provide the reader knowledge of the theory’s limitations.

Factors in the economy affecting the PPP are described below.

4.1.1 Definition

The Swedish economist Gustav Cassel named the theory and stated:

“Our willingness to pay a certain price for foreign money must ultimately and essentially be due to the fact that this money possesses a purchasing power as against commodities and services in that country.

On the other hand, when we offer so and so much of our own money, we are actually offering a purchasing power as against commodities and services in our own country. Our valuation of a foreign currency in terms of our own, therefore, mainly depends on the relative purchasing power of the two currencies in their respective countries.”

Gustav Cassel, economist (1922, pp. 138-39)

The Purchasing Power Parity is an astonishingly simple theory that states that the nominal exchange rate between country A and B must be equal to the purchasing ratio of the price levels between country A and B; implying that one unit of the specific currency in country A has the same purchasing power in a foreign country, country B (Taylor and Taylor 2004). First and foremost, the idea behind the PPP is that one unit of country A’s currency should buy the same mix of goods in country A, as equivalent amount units of country B’s currency should be able to buy in country B. That means

that Country A and B have the same Purchasing Power, at the current exchange rate.

(Taylor and Taylor, 2004) 4.1.2 Law of One Price

The PPP theory was developed from the Law of One Price; but a common mistake is to equate the PPP and the Law of One Price (LOP), but in fact there are significant differences between the two. The LOP is based on the theory of international arbitrage. The LOP states that when the price of internationally traded goods are converted to a common currency, the goods should have the same price in different countries. However, there are some assumptions behind the LOP theory that has been questioned. For example LOP assumes that there is perfect competition in the market, no trading barriers or tariffs, and no transportation costs. In reality, due to the existence of transportations cost as well as trading barriers and tariffs, the prices vary between different countries and the LOP assumptions are questionable (Froot and Rogoff, 1995; Rogoff, 1996).

4.1.3 The Development of the PPP

The Purchasing Power Parity theory has a long and well-debated history that can be can be dated all the way back to the 16^th century, Spain, Europe. During the 19^th century, well-known economists like Ricardo, Marshall, Mill, Goschen et al, tested and developed comparable PPP landscapes (Rogoff, 1996). The theory first got its name back in 1918 by Cassel. Cassel’s ideas about the PPP were developed in a backwash of the World War I and the big collapse of the modern financial system.

Before the Great War, all countries followed the global gold standard; the currencies were bound to gold, and could be converted to gold at fixed prices. This created an exchange rate between countries, which reflected the relative price of gold. After the war started it became quite difficult to maintain the gold standard system, and speculators feared that the countries would devalue their currencies to gain short-term revenues. The gold standard system was abandoned, and many countries let the exchange rate float, or bound it to the dollar (Rogoff, 1996). In this turbulent era, Cassel argued for the accuracy and usage of the PPP theory to explain why the exchange rates should depend on the compared purchasing power between countries.

Since then, the idea of PPP has been discussed and debated in an abundance of reviews and papers among economic researchers all around the globe (Rogoff, 1996).

4.1.4 Critiques of the Applied PPP

There have been several attempts to prove or undermine the theory, and there are arguments both for and against PPP.

There is an easy way to check whether there are deviations from the Purchasing Power Parity, by comparing the difference in price of one McDonald´s Big Mac hamburger around the globe. Named by the hamburger, the Big Mac Index is semi-annually published by the newspaper “The Economist”. The main purpose of the index is to compare prices in a common currency (dollar), to analyse the price differences. The index effectively measures if a currency is over- or undervalued compared to the USD. The assessment is built on the assumption that the currency would be correctly valued, if the international prices of burgers were equivalent dollars (Rogoff, 1996).

This kind of analysis has limitations, and Taylor and Taylor (2004) argued that the approach is a simplification of reality and therefore should not hold. Nevertheless, despite the imitation mentioned, The Big Mac index has been very popular and therefore the Economist created another index describing the price of a tall cup of coffee at Starbucks (The Economist 2013).

The Big Mac- and the Starbucks index are both engaging ways to visualize the exchange rate’s function. Taylor and Taylor (2004) argued that there are several legitimate reasons why the price of a Big Mac or a Starbucks coffee might differ between countries that cannot be explained by the exchange rate. The primary reason is the fact that the economic environment is not perfect (Taylor and Taylor, 2004).

Reasons for international price differences:

 Not easily traded internationally

 Local subjective estimation of the services’ complexity factor affect wages

 Local differences in price level affect costs

None of the factors above are conveniently arbitraged in the international market, and most of the arguments around the PPP are based on the assumption that there are wide international arbitraging. Hence, this indicates that the Big Mac index should be treated with caution when interpreting results (Taylor and Taylor 2004).

As mentioned above, in theory the PPP assumptions should hold when comparing to the “Law of One Price”, but the reality is more complex. If the price of a product differs in different regions, a person could make a totally riskless profit by shipping the good from area A, where the price of the good is low, to area B where the price is relatively higher. This statement ignores the fact that there are transaction costs in the market, which hinder the good from being moved from A to B and lower the price.

Furthermore, one of the PPP theory’s limitations is the definition of the market basket in order to compare goods internationally. The taste and preference of goods between countries differ, and it is therefore not accurate to use the same basket for every country. Regardless of the preferences, there are also difficulties in purchasing exactly the same goods in different countries due to availability. However, if there was an accurate market basket of traded goods, in a perfect market without transaction costs, the Law of One Price would imply that the PPP theory of an adjusted exchange rate should hold (Rogoff, 1996). As mentioned, there are strong objections why the relationship between the LOP and PPP cannot hold. Different transaction costs disaffirm the theory; e.g. transportation cost, trading barriers, tariffs, taxes etc. (Engel and Rogers 1996).

Further, there are goods that are not traded between all countries in the world;

consequently the weights of those goods are not going to be equal in the countries’

market baskets. Moreover, observations have also shown that countries tend to create differentiated goods instead of substitutes. Some of these problems, however, could be addressed by using more accurate statistic data (Engel, 1996).

Taylor and Taylor (2004) argue that since the PPP is based on the traded goods, a more appropriate test is the PPI (producer price indices) in comparison to the CPI (consumer price indices). Taylor and Taylor (2004) claim that the PPI usually reflect the tradeable manufacturing goods, while the CPI tends to reflect more non-tradeable goods. By using PPI, the analysis would be more accurate according to Taylor and Taylor (2004).

In this thesis the aim is to test whether the Purchasing Power Parity holds for services internationally, or if it is too simplified and the problems mentioned above are actual issues.

5. Methodology

In the regression tests, panel data is used. The regressions are run with data for macro variables clustered by country. The country clusters are independent, but the factors contained in the cluster are dependent. In the regressions, Huber (1967) –White (1980), standard errors are used to adjust for clusters. xtreg matches the regression models to the panel data for both random and fixed effect. The Cost of Living data is unbalanced, since each country is not observed every year. The Engineer and Fitter data is balanced, since each country is observed every year.

The data is collected for the years 2001-2011. Accounting data, economic data and exchange rate data are collected from Thomson’s DataStream. The Cost of Living index is collected from U.S. Department of State, and the Big Mac Index is retrieved from The Economist. Agriculture data, unemployment data and poverty $ 2 are collected from World Bank. The Cost for Fitters and Engineers are collected from Flexlink’s financial statement. In the world (general measure) regression the dependent variable is Cost of Living, tested for 9 macro-variables in 7 models. The sample consists of 188 countries and the number of years is 1880. For the regression based on Flexlink’s cost of Fitters and Engineers as dependent variables, the tested macro- variables are 9 in 7 models for Fitters, and 9 in 5 models for Engineers. The sample consists of 19 Flexlink Sales Units and the number of years is 190.

5.1 Macro Variables on Pricing

In order to investigate the macro factors that have an impact on service pricing, the variables are first discussed below. The variables are seen as representative of the macro- and micro economic factors. There are two types of variables examined;

Dependent and Independent.

5.1.1 Dependent Variables

The dependent variables examined are Cost of Living Index and the hour rate for Fitters/Engineers at Flexlink.

5.1.1.1 Cost of Living Index (COL_IN)

The Cost of Living Index has Washington DC as base (Washington DC=100). The index compares the costs in Dollars ($) of buying representative goods and services (excluding education and housing) in foreign countries with the cost of equivalent goods in the DC area. The index excludes all kind of extravagant goods and services, and is based on the average American consumption behaviour. The Cost of Living Index has limitations in measuring and comparing costs differences between countries.

The Index focuses on the comparison of American families and has DC=100 every year, which makes it impossible to use it for measuring cost changes over time in a foreign country. The Cost of Living Index is used in this thesis to examine if there are any differences between countries and if there are any similarities between Cost of Living and Cost of Engineers and Cost of Fitters at Flexlink (U.S. Department of State 2011).

5.1.2.2 Cost of Engineer

The Cost of Engineer is the average hourly rate for an Engineer at Flexlink, in 19 different locations around the globe. The variable includes Sales and Administrations costs, wages, rents, and social expenses etc.

5.1.2.3 Cost of Fitter

The Cost of Fitter is the average hourly rate for a Fitter at Flexlink, in 19 different locations around the globe. The variable includes Sales and Administrations costs, wages, rents, and social expenses etc.

5.2.3 Independent Variables 5.2.3.1 The Big Mac Index

The Big Mac Index (BMI) was created in 1986 by the magazine The Economist. The index is based on the idea that the exchange rates are reflected in the price levels, when two countries are compared, known as the Purchasing Power Parity (PPP) (Clements, Lan, & Seah, 2012). The Big Mac index is therefore an indication if the currency is under- or overvalued (Clements, Lan, & Seah, 2012). The Big Mac works very well for representing a bundle of goods described by the PPP, since the hamburger consists of different raw materials e.g. salad, beef, and cheese and is therefore a basket on its own (Dewhurst, 2003). McDonald’s is a big brand and

restaurants with a standardized menu are located in 119 countries for comparison (McDonalds Corp., 2013).

The price of a burger is not only dependent on the input to create a hamburger, but also on wages and the cost of hiring a venue (The Economist, 2009). This view is strengthen by Parsley and Wei (2007) whom claim that there are very little difference between the ingredients that makes a hamburger, and therefore the BMI works very well as a determinant of the correct price for tradable and non-tradable goods. From these close to perfect conditions the Big Mac is a helpful tool to create the absolute PPP. To investigate whether the PPP theory holds, and if prices can be standardized, BMI is a good measure of the current situation worldwide and therefore included in the test.

5.2.3.2 GDP-Deflator

The GDP-deflator measures inflation in a country by comparing the current local currency to the constant local currency¹. The deflator is obtained by looking at the price level from a yearly basket of all new in-country produced goods and services, and comparing to a base level to assess the index². The nominal GDP evaluates the economic worth of transactions in prices for the given period, and thus measures both volume and price. The constant GDP measures the exact change of volume for goods and services for the specific time (Chowdhury, 2008).

GDP−Deflator =Nominal GDP Real GDP ×100

It is of importance to measure inflation in comparison to other countries as the strength of the currency decides the purchasing power. The advantage of the GDP- deflator is that it covers a broad spectrum of the economy and incorporates the changes that are subject to both consumption and the monetary policy (Chowdhury, 2008). In comparison to BMI the GDP-deflator is a more widely accepted economic variable to measure inflation, and thus included to test wheter the PPP-theory holds in addition to BMI.

1 (http://data.worldbank.org/indicator/NY.GDP.DEFL.KD.ZG).

2 (https://www.boundless.com/economics/measuring-nation-s-output-and-income/gdp-trends- using-real-and-nominal-values/calculating-gdp-deflator/).

5.2.3.3 Poverty lower limit 2$

The minimum income of $2 a day to survive, have been drawn as a guideline worldwide (World Bank). There are many negative effects to low income. The poor may not be able to seek medical care or buy enough food, and the result is decreased productivity in the country as the society cannot make use of the citizens’ maximal capacity. Those who are poor tend to have less education or be illiterate that results in worse jobs that thus have lower wages (UN, 2013).

The poverty line is a generalization of the minimum required income in order to stay alive, but very useful in determining the situation in a country and how poverty affects PPP regarding pricing. In countries with a high proportion of poor, the price level after being adjusted for exchange rates should give the same purchasing power if the poverty variable has no effect on pricing, but it can be assumed that the prices are lower and not in accordance with the PPP theory. The poverty variable is used in this thesis to test if it is possible to find any significant impact between how poor a country is, the Cost of Living and service pricing.

5.2.3.4 Agriculture

In general, as a society develops, the amount of people working in agriculture decreases. A majority (75%) of the poorest in the world live in rural areas, and it is often far to the nearest town. Even in developed countries 45% of the poor live in rural areas. The only way to make a living is by growing crops or raising livestock, and therefore agriculture make up 70% of their income (World Bank, 2013). For the absolute poor, it is of particular importance since agriculture is the best way to improve their situation due to the fact that they often lack education. It is also positive for the poorer countries, where food needs are secured, and surplus harvest can be sold and exported, leading to growth for the entire country (World Bank, 2013). Generally, where there are more poor people, there are more agriculture and this should affect prices to be lower, even after adjusting for exchange rates and proof against the standardization of prices. Lower prices should also have, at lest some impact on the Cost of Living and it is therefore of great interest to test this variable in the regression models.

5.2.3.5 GDP/Capita

GDP stands for gross domestic product and is the sum of all economic activity in a

investments and adjusted for export minus import. A common welfare measure is GDP/capita that describes the amount of GDP per person living in the country. The higher this ratio is the better of is the country economically. There have been a lot of critiques against the validity of the GDP/capita. Kennedy (1968) argues that it is purely an economical average and does not actually measures the wellbeing of people e.g.

education, healthcare. GDP/Capita measures how developed/effective a specific country is, when it comes to output per person. It is therefore interesting to examine how big impact the factor has together with the Agriculture index and the Poverty index on Cost of Living and service pricing, to see if there are any linkages between the effectiveness of a country and the service pricing and Cost of Living.

5.2.3.6 Wage

In developed countries, the wage level is generally higher since people in those countries are more educated and that pushes up wages. The effects are that a higher wage level creates a beneficial situation for abroad travels and import. In developing countries the price level is usually lower, which makes the lower wages purchasing power, in-country, more equal to the developed countries. The factor is included in the analysis because wage is a main component in service pricing. In order to test if this is true, the variable is used. Wage in comparison to the price level in the country is a significant factor after adjusting for exchange rate, for PPP to hold. If the wage to price level ratio differs more than the exchange ratio, it is proof against PPP.

5.2.3.7 Unemployment

The higher the unemployment, the more costs to the society due to social security. If less people are employed, the side effect is that consumption will decline causing hardship for the entire economy. If employment opportunities are scarce, less people will invest in education and businesses (Investopedia, 2013). A high unemployment percentage increases the supply of workers that competes for available jobs, and has a negative effect on wage. Lower wages cause the cost of service to decrease and is therefore important to test.

5.2.3.8 Exchange rate

A strong currency makes import cheaper, but on the contrary makes export harder since the goods will be perceived as more expensive by other countries. The exchange rate ratio, for that reason, has a strong impact on the trading situation. The exchange

rate should be a significant factor according to the PPP theory, and is therefore important in the tests.

5.2.3.9 Service employed

Too estimate the value of the service sector; it is useful to know how many percent of the total workforce is employed in the service industry. This shows the importance of this sector to the country. The factor is included to examine if the service-employed level has a significant effect on both the Cost of Living and the service pricing.

5.3 The Regression Model

In the previous part of the paper, the PPP theory was presented and the macro variables of interest were introduced. To be able to answer the research questions, statistical tests are required that analyze the economic theory and the variables. The selected dependent variables are labor cost and Cost of Engineer/Fitter, and it is going to be examined whether there are any relationships between the variables and the independent macro variables. An estimation of the variables betas together with a hypothesis test will determine the statistical and economic significance. Panel data will be used in the regressions.

There will be three parts to the test since the aim is to investigate the variables impact on:

1) A general measure (worldwide) 2) Engineers

3) Fitters

The three models of costs can be compared and discussed, for a comprehensive view and conclusion.

Cost of Living Index=β₀+β₁∗GDP Deflator +β₂∗Big Mac Index (PPP )+β₃∗Unemployment (%)+β₄∗Agriculture (%)+β₅∗GDP/CAP+β₆∗Exchange rate+β₇∗Wage_World+β₈∗Service employed (%)+ β₉∗Poverty $ 2 cos T =β +β ∗GDP Deflator+ β ∗Big Mac Index ( PPP)+β ∗Unemployment (%)+ β ∗Agriculture (%)+β ∗GD P/CAP+ β∗Exchange rate+β∗Wag e +β∗Service employed (%)+β∗Poverty $ 2

cos T_FIT=β₀+β₁∗GDP Deflator+ β₂∗Big Mac Index (PPP)+β₃∗U nemployment (%)+β₄∗Agriculture ( %)+β₅∗GDP /CAP +β₆∗Exchange rate+β₇∗Wag eper country+β₈∗Service employed (%)+β₉∗Poverty $ 2

5.4 Hypothesis

Hypothesis 1: Cost of Living H₀: β₁+β₂+...+ β₉=0 H₁: β₁+β₂+...+β₉≠ 0

H₀: The macro factors have no effect on COL

H₁: At least one of the macro factors has an effect on COL Hypothesis 2: Cost of Engineer

H₀: β₁+β₂+...+ β₉=0 H₁: β₁+β₂+...+β₉≠ 0

H₀: The macro factors have no effect on COST_ENG

H₁: At least one of the macro factors has an effect on COST_ENG Hypothesis 3: Cost of Fitter

H₀: β₁+β₂+...+ β₉=0 H₁: β₁+β₂+...+β₉≠ 0

H₀: The macro factors have no effect on COST_FIT

H₁: At least one of the macro factors has an effect on COST_FIT

6. EMPIRICAL RESULTS

6.1 Correlation Results

Table 1 - Correlation between the macro factors in the World regression models; calculations below are based on data observations in 188 countries. The values are adjusted for Robustness. The table shows strong correlation between AW_WAGE and GDP/CAP (0,9332) and the two should therefore not be tested in the same regression model.

AGRI_ UNEMP_ BMI_PPP GDP/CAP GDP_DEF COL_IN EX_RATE SERV_% POVU_2_ AW_WAGE

AGRI_% 1.0000

UNEMP_% -0.1121 1.0000

BMI_PPP 0.2596 0.0686 1.0000

GDPCAP -0.5175 -0.2308 -0.1957 1.0000

GDP_DEF -0.0966 0.0055 -0.0786 0.1126 1.0000

COL_IN -0.1909 -0.0196 -0.1929 0.5239 0.0906 1.0000

EX_RATE -0.3474 0.2064 -0.2102 0.2775 0.0232 0.2691 1.0000

Serv_% -0.5472 0.0604 -0.2977 0.3957 0.0110 0.1819 0.3457 1.0000

POVU_2_ 0.7741 -0.2132 0.2967 -0.7154 -0.0521 0.0485 -0.3452 -0.5578 1.0000

AW_WAGE -0.5428 -0.1257 -0.1368 0.9332 -0.1300 0.7735 0.2102 0.3777 -0.4478 1.0000

AGRI_% = Percentage of the population working in the agriculture sector UNEMP_% = Percentage of the population unemployed

BMI_PPP = Big Mac index – Purchasing Power Parity adjusted GDP/CAP = The GDP/ Capita in Dollar

GDP_DEF = The GDP Deflator

COL_IN = Cost of Living index, with USA as base country EX_RATE = Exchange rate, with SEK as base currency

POVU_2_% = Percentage of the population earning less than 2 dollars/day SERV_% = Service sectors percentage part of the GDP

AV_WAGE = Average wage/hour

Table 2 - Correlation between the macro factors in the Flexlink regression models; calculations below are based on the 19 target countries. The values are adjusted for Robustness. The table shows high correlation between the independent variables POVU_2_ and AGRI_, and the two should therefore not be tested in the same regression models.

AGRI_ UNEMP_ BMI_PPP GDPCAP GDP_DEF AV_WAGE EX_RATE ENG_COST FIT_COST POVU_2_ SERV_

AGRI_% 1.0000

UNEMP_ -0.2606 1.0000

BMI_PPP 0.3961 0.3729 1.0000

GDP/CAP -0.5854 0.1708 -0.2356 1.0000

GDP_DEF 0.0553 0.0142 -0.0475 -0.1924 1.0000

AV_WAGE -0.7406 0.1139 -0.2981 0.5044 -0.3334 1.0000

EX_RATE -0.0421 -0.2340 -0.2862 0.0680 0.1984 0.0881 1.0000

ENG_COST -0.6382 0.0420 -0.2918 0.4465 -0.3226 0.9333 0.1232 1.0000

FIT_COST -0.8740 0.1163 -0.3797 0.4941 -0.1295 0.9298 0.1159 0.9492 1.0000

POVU_2_ 0.9168 -0.1516 0.5278 -0.5646 0.0487 -0.5572 -0.0291 -0.5186 -0.7258 1.0000

SERV_% -0.8478 0.1152 -0.5336 0.4259 -0.0611 0.7893 0.2570 0.7261 0.8952 -0.7094 1.0000

AGRI_% = Value added as a percentage of GDP UNEMP_% = Percentage of the population unemployed

BMI_PPP = Big Mac index – Purchasing Power Parity adjusted GDP/CAP = The GDP/ Capita in Dollar

GDP_DEF = The GDP Deflator AV_WAGE = Average wage

EX_RATE = Exchange rate, with SEK as base currency ENG_COST = Average Engineer cost/hour at Flexlink FIT_COST = Average Fitter cost/hour at Flexlink

POVU_2_% = Percentage of the population earning less than 2 dollars/day SERV_% = Service sectors percentage part of the GDP

6.2 Regression Results

Table 3. Regressions with Cost Of Living as dependent variable

The sample consists of 188 countries 2001-2011. The number of years is 1880. Accounting data, economic data and exchange rate data are collected from Thomson’s DataStream service.

The Big Mac Index is collected from The Economist. Agriculture data, unemployment data and poverty $ 2 are collected from World Bank. The dependent variable is Cost of Living (COL_Index), collected from the U.S. Department of State, Bureau of Administration. We average all variables and run the regression on all 188 observations when enough data matching the observations are available. AGRI_% is equal to the value added as a percentage of GDP (in a specific country). UNEMP_% is equal to the percentage of the population (in a specific country) unemployed. EX_RATE is equal to the exchange rate with Swedish Krona as base currency. BMI_PPP is equal to the PPP adjusted Big Mac index with USA as base country. GDP/CAP is equal to GDP per capita measured in Swedish Krona. AW_WAGE is the average total wage in the country. GDP_DEF is almost the same as BMI but with adjusted goods baskets depending on the countries preferences and supply. SERV_% is equal to the percentage working in the service sector. POVU_$2_% is equal to the percentage earning less than 2 dollars a day. The confidence interval is set to 95 %. The degrees of freedom are 1870.

The bold numbers are the coefficients, and the values in parentheses are the p-values. In the regressions, Huber (1967) –White (1980), standard errors are used to adjust for clusters.

COST OF

LIVING INDEX M1 M2 M3 M4 M5 M6 M7

AGRI_ % -2.516769 .2560752 -2.548741 -4.080629

(0.000) (0.326) (0.000) (0.000)

UNEMP_ % -.3172955

(0.633)

BMI_PPP -.0002274 .0157492 -.0000752 .0000175

(0.916) (0.410) (0.965) (0.995)

GDP/CAP .0026859 .0005467

(0.000) (0.004)

GDP_DEF .4344633 .4122582 .3653433

(0.141) (0.319) (0.188)

AV_WAGE .6879482 1.063822 1.015382

(0.000) (0.000) (0.000)

EX_RATE .3390573

(0.337)

SERV_% .2551647

(0.659)

POVU_$2_% -.2782521 .199643

(0.122) (0.061)

R2 0.6934 0.1619 0.1921 0.6133 0.5921 0.6446 0.5465

N 370 380 930 350 360 470 490

Summary: Strong correlation between the variables AW_WAGE and GDP/CAP (0,9332), and the two are therefore not tested in the same regression model. The models show that AGRI_%, GDP/CAP and AV_WAGE have significant effects on Cost of Living.

Table 4. Regressions with Engineer Cost in Flexlink as dependent variable

The sample consists of 19 Flexlink Sales Units 2001-2011. The number of years is 190.

Accounting data, economic data and exchange rate data are collected from Thomson’s DataStream service. The Big Mac Index is collected from The Economist. Agriculture data, unemployment data and poverty $ 2 are collected from World Bank. Wage information from Flexlink is collected manually from each Sales Unit’s Finance Manager. The dependent variable is Cost of an Engineer hour at Flexlink (COST_ENG). We average all variables and run the regression on all or 19 observations. AGRI_% is equal to the value added as a percentage of GDP (in a specific country). UNEMP_% is equal to the percentage of the population (in a specific country) unemployed. EX_RATE is equal to the exchange rate with Swedish Krona as base currency. BMI_PPP is equal to the PPP adjusted Big Mac index with USA as base country. GDP/CAP is equal to GDP per capita measured in Swedish Krona.

AW_WAGE is the average total wage in the country. GDP_DEF is almost the same as BMI but with adjusted goods baskets depending on the countries preferences and supply. SERV_%

is equal to the percentage working in the service sector. POVU_$2_% is equal to the percentage earning less than 2 dollars a day. The confidence interval is set to 95 %. The degrees of freedom are 180. The bold numbers are the coefficients, and the values in parentheses are the p-values. In the regressions, Huber (1967) –White (1980), standard errors are used to adjust for clusters.

COST_ENG M1 M2 M3 M4 M5

AGRI_ % -21.63403

(0.003)

UNEMP_ % -2.383902

(0.894)

BMI_PPP -.0004061

(0.991)

GDP/CAP -.0000599 .0000772 .0003627

(0.804) (0.842) (0.352)

GDP_DEF -.0425837 -.868106

(0.811) (0.01)

AV_WAGE 2.211463 (0.000)

EX_RATE -2.609293

(0.620)

SERV_% 14.5916 13.74858

(0.000) (0.000)

POVU_$2_% .1550995 -5.550969 1.042789

(0.905) (0.091) (0.554)

R2 0.8720 0.5316 0.4912 0.2703 0.5542

N 190 190 190 190 190

Summary: Strong correlation between the variables POVU_$2_% and AGRI_% (0.9168), and the two are therefore not tested in the same regression model. The models show that AGRI_%, SERV_% and AV_WAGE have significant effects on Cost of Engineer.