The Winning of Natural Gas and Its Effect on Real Estate Value: Exploring the Design of a New Monitoring System

(1)

University of Groningen

Faculty of Economics and Business

MSc Technology and Operations Management

The Winning of Natural Gas and Its Effect on Real

Estate Value: Exploring the Design of a New

Monitoring System

By MARCEL BAKKER Ernst Casimirlaan 54 9717AX Groningen M.Bakker.34@student.rug.nl Student number: S2233045 23-01-2015

(2)

1

ABSTRACT

The number of earthquakes in the region of Groningen due to gas winning is increasing. At the point of writing this, an earthquake of 2.8 on the scale of Richter hit Groningen. These earthquakes can have a large impact on properties in the risk area and therefore they can affect the value of properties in the region of Groningen. The University of Groningen therefore proposed to design a system that can monitor the effect of the winning of gas on real estate value. This research will contribute to the monitoring system by analysing the goals all stakeholders can have in a monitoring system and by proposing an improved model of the current claims settlement process used by the Nederlandse Aardolie Maatschappij (NAM). This improved model can be used in future research such that it can be used to build a monitoring system.

In order to determine the stakeholders, their goals and the Critical Success Factors (CSF’s) in solving these goals, interviews were conducted with the stakeholders. The stakeholders were identified as the Ministry of Economic Affairs, the NAM, the Province of Groningen, the municipalities, the residents, who are represented by the Vereniging Eigen Huis (VEH) and Groninger Bodem Beweging (GBB), and lastly the realtors who are represented by the Nederlandse Vereniging van Makelaars (NVM). It must be clear to the reader that the stakeholders are far from agreeing with each other about how to measure the influence of gas winning on real estate value. Ortec Finance on behalf of the Ministry of Economic Affairs has performed research. This quarterly repeating research mostly concludes that there is no significantly lower real estate value in the risk area. The NAM also has its calculation method to determine whether there is a loss in value of an individual property due to the winning of gas. The method used by Ortec Finance and the NAM are both criticised by the other stakeholders as they conclude both that there is little to no loss in value due to the winning of gas.

The results show that not a single stakeholder except for the NAM and Ministry of Economic Affairs have a calculation method to determine the influence of the winning of gas on real estate value. This thesis will show an improved version of the method currently used by the NAM. This can be used in future research in order to build a monitoring system.

(3)

2

PREFACE

This thesis is written as a final project for finishing my Master degree program Technology and Operations Management. It was written for the University of Groningen and was assessed by dr. H. Balsters and prof. dr. J.P. Elhorst. I would like to thank both gentlemen for their excellent guidance. Despite that there was no contact with stakeholders at the time this project started, we still managed to produce results that can be helpful to a lot of stakeholders.

Also, I would like to thank all persons I have interviewed during this research for their cooperation and time they made available. Without them it would not have been possible to write this thesis.

(6)

5

1 INTRODUCTION

In 1963, the Nederlandse Aardolie Maatschappij (NAM, in English: Dutch oil company) started winning gas in the northern part of the Netherlands (KNMI 2013). With 900 km2 the Groningen gas field is one of the largest in the world. From 1963 to present, revenues totalled well over 200 billion Euros and thus is of great importance to the state of the Netherlands (NAM 2014a). Up to 1986 the winning of gas had no effect in terms of earthquakes. However, from 1986 to present there have been measured approximately 1000 earthquakes by the Koninklijk Nederlands Meteorologisch Instituut (KNMI, in English: Royal Dutch Meteorology Institute), which were the result of the winning of gas (KNMI 2013). These earthquakes have a variety of impacts on the residents and their properties in the risk area including the decline in value of properties (Van der Voort & Vanclay 2015). The impact of these earthquakes is expected to become larger as the yearly number of earthquakes and its strengths will rise in the coming years thereby increasing the urgency to measure the influence of gas winning on real estate values (Muntendam-Bos & de Waal 2013).

The University of Groningen would like to address this issue by studying the impact of earthquakes on the living conditions and well-being of homeowners in the risk area (de Kam & Elhorst 2014). In order to do so, the University of Groningen requested co-operation and collaboration of the NAM, the Ministry of Economic Affairs and the Province of Groningen, also known as the dialoogtafel, from now on indicated as dialogue table. The University of Groningen would like to co-operate and collaborate with these parties to see which sources of information are needed to build an information system in which the effect of the winning of natural gas on real estate values is monitored. The dialogue table has not yet approved the request for co-operation and collaboration. Therefore, this research was conducted without the collaboration of the dialogue table but with collaboration of various stakeholders, which will be identified in this research.

(7)

6 the third quarter of 2013. The reports do conclude that the sales process of real estate in the risk area is more difficult (Francke & Lee 2014a).

To analyse whether a difference in value of real estate is present in the risk area, Francke & Lee (2013) compared the selling price trend of real estate over time in the risk area and predetermined reference areas. These reference areas are nearby areas comparable with the risk area, given social-economical and demographical factors. The reference areas are subdivided into municipalities (Francke & Lee 2013). Two methods were used to calculate real estate value and to analyse whether a difference in value due to gas winning is present. The first method is the hedonic pricing method. This method estimates the value of environmental advantages or disadvantages from the prices of related market transactions and takes into account multiple characteristics of the property (Tyrvainen & Miettinen 2000). For this method, data of the Nederlandse Vereniging van Makelaars (NVM, in English: Dutch Realtors Association) is used. The second method is the repeat-sales method. This method uses sales prices for the same property at different points in time to calculate the difference in value in a given area (Clapp & Giaccotto 1992). For this method, data of the Kadaster (Land registry) is used. The characteristics of the data used are summed up for each data source (see Table 1.1)

Table 1.1: Characteristics of NVM and Kadaster database

Characteristics NVM Kadaster

Available from 1985 1993

Percentage of transactions 70% 100%

Financial characteristics - Asking price - Selling price

- Selling price

Property specific characteristics - Size of property - Land size - Type of property - Year built - State of maintenance - Additional features - Type of property - Land size

(8)

7 Several other limitations can be identified in the way Ortec Finance conducted its research. De Kam & Elhorst (2014) summed up four more limitations. Firstly, the defined risk and reference areas are incomplete as it is possible for earthquakes to occur in areas in which they have not been yet. This is also the conclusion of Muntendam-Bos & de Waal (2013), who state that the size of the area where earthquakes are observed is increasing. Secondly, a number of characteristics, which are essential, are not used in the hedonic pricing method, such as the type of subsoil or the presence of earthquake proof renovations (see Table 1.1). Tsai (1969) agrees to this as he concludes that the type of subsoil has an effect on the way an earthquake behaves and thus the extent to which an earthquake creates damage to a property. Thirdly, Ortec Finance uses data from the NVM for the hedonic pricing method (Francke & Lee 2013). The advantage of this data is its extensiveness, however, it only contains 70% of all transactions since 1993 thus leaving out an average of 30% (see Table 1.1). Fourthly, Ortec Finance did not make its used database publicly available. As a result of this, it is not possible for others to replicate the research.

The University of Groningen would like to contribute to resolve the limitations of the research of Ortec Finance by proposing to design a monitoring system for real estate value. This system will be capable of measuring the influence of gas winning in the Groningen region on real estate value by combining various data sources. This research will contribute to the monitoring system by analysing the goals all stakeholders can have in a monitoring system and by proposing an improved model of the current claims settlement process used by the NAM. This improved model can be used in future research such that it can be used to build a monitoring system. The research question can be defined as the following:

How can the current claims settlement process of the NAM be improved such that it can be used as a basis for a system that can monitor the effect of natural gas winning on the value of

real estate in the Province of Groningen?

(9)

8

2 RESEARCH DESIGN

The type of problem being investigated can be defined as a knowledge problem (Wieringa 2007). As such, the structure of this report is based on his description. This chapter describes the method that has been used as structure for performing the research. Also, the different models and methods that have been used for this research and its results will be described.

2.1 Regulative cycle by Van Strien

Strien (1997) developed the regulative cycle, which can be used as structure of research in design science. This cycle consists of five steps, of which three will be treated. These phases will be explained in this section. The cycle can be seen in Figure 2-1. The questions that need to be answered are given for each phase, and are based on Balsters (2014b).

Figure 2-1: Regulative cycle by Van Strien

Design Problem - The first step is referred to as Design Problem. This step consists of three main questions that will be answered:

1. Who are the stakeholders? 2. What are their goals?

3. What are the Critical Success Factors (CSF’s) for each goal? These questions will be answered in chapter 3.

Diagnosis/Analysis - Once the stakeholders, their goals and their CSF’s have been identified, the diagnosis/analysis can start. For this, two main questions will be answered.

1. What are possible causes of the difficulty of resolving a CSF? 2. In what order do the CSF’s need to be treated?

This phase will be treated with the verified goals and CSF’s.

(10)

9 1. Which solution alternatives are available?

2. Is it possible to gather old solutions to build a new? 3. Is it necessary to invent a new solution?

Additionally, the Hour Glass Model has been used to come to a proposed solution. This model will be explained in section 2.2. Information regarding at which organisations data can be collected to fill the model is described in Appendix D.

Implementation - This phase will not be treated in this research as this research only focuses on building a proposed model and thus not treating the implementation.

Validation - Due to the limited amount of time available, it was not possible to validate the solution. Therefore, the proposed model must be further worked out and validated in future research.

2.2 Hourglass Model

A relevant model that has been partially used in this research is the Hour Glass Model. This model by Balsters (2013a) is made to link individual data sources to each other in order to let different apps extract information from these sources. The model, which can be seen in Figure 2-2, consists of a lower and an upper part. The lower part contains all available data sources. The upper part contains a number of apps. These apps can give usable data as output. The apps are linked to a data federation. This data federation contains the ideal data, which is necessary for the apps. The critical part of the Hour Glass model is to link the actual data sources to the data federation such that the apps can output usable data. The data federation does not necessarily copy the data to its own database, it just links together the different data sources available (Haas et al. 2002).

(11)

10 Figure 2-2: Hour Glass Model (Balsters 2013a)

2.3 Academic relevance

(12)

11

3 DESIGN PROBLEM

This chapter describes the current difficulties in the context of the problem. Firstly, an introduction to the relevant difficulties in the context is given after which the method that has been used to verify stakeholders’ goals will be given. Lastly, the identified stakeholders, their goals and CSF’s will be given. Note that the verified goals and CSF’s are based on an earlier set of expected goals and CSF’s, which can be found in Appendix B.

3.1 Introduction to the context

The opinions and goals of the stakeholders seem far from aligned with each other. This becomes very clear when looking at different reports in the media. Residents complain that the earthquakes have caused severe structural damage to their properties and that the NAM is offering too low compensations to repair the damage (RTV Noord 2014a). This is also the conclusion of the Vereniging Eigen Huis (VEH, in English: Private Property Association), an organisation that represents the interests of property owners. They state that the methods used by the NAM to evaluate the damage are insufficient (Telegraaf 2014). As stated, the government hired Ortec Finance to determine the possible decline in value of properties in the risk area. Statements based on this research from the Ministry of Economic Affairs puzzle realtors and interest associations. They question the statement from the Ministry of Economic Affairs that there is no decline in value of properties at this moment (RTV Noord 2014c). The government also seems divided when it comes to the problems caused by earthquakes. According to Eric Smaling, a member of the Dutch government, the Minister of Economic Affairs should empathize more with the residents of the risk area (RTV Noord 2014b). The local government also questions the current way of working related to the winning of gas. The Board of Mayor and Aldermen of Groningen state that the limits have been reached when it comes to gas winning. They fear that on-going construction projects will become a lot more expensive when they need to be earthquake proof (Dagblad van het Noorden 2014a). Between the different municipalities there is also a difference in opinions. Municipalities in the earthquake risk area are not willing to share subsidies made available to improve properties with other non-earthquake affected municipalities (Dagblad van het Noorden 2014b). Also, the dialogue table, which was introduced to take care of the interests of the stakeholders of this problem, is losing its trust from the residents (Havermans 2014).

(13)

12 how to solve this problem. This can cause difficulties in this research. The large amount of news related to the problems of gas winning also shows that the context keeps on changing. New ways of preventing earthquakes by injecting water or nitrogen could have a high impact on the need for a monitoring system and the content of the system (Overduin 2014). Also, new stakeholders could get involved in the process in such a way that the monitoring system needs to be adapted to it. The VEH for example got involved in the problem more recently by criticising the rules the NAM uses for the claims settlement process in case of structural damage (Telegraaf 2014).

3.2 Identified stakeholders

Prior to determining the stakeholders, the definition of a stakeholder must be clear. Mitchell et al. (1997) define a stakeholder as any group or individual who can affect or is affected by the research objective. This definition will be used when identifying the stakeholders. Based on the work of De Kam & Elhorst (2014) and an interview with prof. dr. ir. G.R.W. de Kam, several key stakeholders have been identified. For each of the stakeholders an explanation will be given.

Ministry of Economic Affairs - Gas revenues are in total over 200 billion euros, thus making the winning of gas an important source of income for the Netherlands. If the outcome of this research would mean that more compensation is given to the residents, this money must be deducted from the gas winning revenue thereby lowering the income for the state of the Netherlands. Therefore, the Ministry of Economic Affairs is an important stakeholder of this research.

NAM - A key stakeholder in this research is the NAM as it is responsible for the winning of gas in the region of Groningen. The NAM is eventually the organisation that will compensate those who have damage due to the earthquakes. They also have data regarding the number of complaints about earthquakes or damage caused by earthquakes.

Local authorities - Two main local authorities can be identified as a stakeholder, namely the Province of Groningen and the municipalities in the risk area.

(14)

13 Municipalities in the risk area - The municipalities in the risk area are the closest government body to the residents of the risk area. They also must take care of the interest of their residents and the municipalities also have the Waarde Onroerende Zaken (WOZ, in English: property value) value of the properties in their area. This WOZ value can possibly be used for the monitoring system.

Residents in the risk area - The stakeholder who is affected directly are the residents in the risk area. Among others, the Groninger Bodem Beweging (GBB, in English: Groninger Soil Movement) and the VEH represent the residents in the risk area. Therefore, they are included in this research.

VEH – The VEH represents property owners in the Netherlands. They stand for the interests of property owners and thus also have an interest in a monitoring system. GBB - The GBB represents the interests of her members who are affected by damage due to the gas winning in Groningen (Groninger Bodem Beweging 2013). Therefore, the members of the GBB have an interest in a monitoring system for the value of properties.

Realtors - The income of realtors depends on the number of properties they sell (Benjamin et al. 2002). The conclusion of Francke & Lee (2014a) is that the sales process in the risk area is more difficult. Therefore, the realtors in the risk area can expect a decline in income due to the winning of gas. Therefore, realtors can also be identified as stakeholders. An organisation that represents realtors is the NVM. They can therefore be seen as a stakeholder.

3.3 Method for verifying stakeholders’ goals and CSF’s

Expected goals and CSF’s for a monitoring system have been determined for all stakeholders, they can be found in Appendix B. These goals needed to be verified in order to be used. The goals and CSF’s were verified by analysing how each stakeholder sees the process how residents can claim a loss in value due to the winning of gas at the NAM. More specifically, how the change in value of real estate in the risk area is calculated. Besides this, it is also analysed what goals all stakeholders can have in a system that can monitor the real estate value.

(15)

14 of real estate value on the reports made by Ortec Finance. For stakeholders whose information was not publicly available, interviews have been conducted. These interviews have been done by phone or in person. Table 3.1 summarises which persons have been interviewed for each stakeholder. Note that the NVM referred to the NAM when a request for an interview was done. Therefore, the data from the interview conducted with the NAM is used for the NVM.

Table 3.1: Contacting stakeholders Stakeholder Organisation to contact Contact Ministry of

Economic Affairs

None The necessary information can be extracted from the report by Ortec Finance

NAM NAM An interview has been conducted with Goffe

Venema

Local Authorities Province An interview has been conducted with, Jeroen Bakker, Huub Hansen and Werna Udding

Local Authorities Municipalities A telephonic interview has been conducted with Jinko Rots from the municipality of Loppersum and Paul van den Berg from the municipality of Eemsmond

Residents VEH A telephonic interview has been conducted

with Steven Wayenberg

Residents GBB An interview has been conducted with

Lambert de Bont

Realtors NVM Frank Harleman from the NVM referred to

the NAM as they can tell how properties are appraised according to the NVM. Therefore, the data from the interview with Goffe Venema is used for the NVM.

In order to ensure the correct information was gathered from the interviews, an interview protocol has been made. This protocol can be found in Appendix A. All interviews are described in detail in Dutch and are therefore not included in this report, they can be found in the document “Thesis Marcel Bakker – Interviews”. A more brief description of all interviews is given in the following paragraphs followed by the verified goals. A summary of all goals and CSF’s is given in Table 3.2.

3.4 Verified Goals and CSF’s – Ministry of Economic Affairs

(16)

15 3.3, the calculation method used by Ortec Finance will be analysed in order to determine the goals and CSF’s of the Ministry of Economic Affairs.

The latest report published by Ortec Finance contains an update of real estate value for the first two quarters of 2014 (Francke & Lee 2014b). All updates use the same calculation method for determining the value of real estate and also use the same reference areas to compare price fluctuations with. The first step in the research of Ortec Finance was the definition of reference areas with which the risk area is compared. Two main reference areas were defined, the first is neighbouring the affected municipalities and the second is neighbouring the first main reference areas. The reference areas were chosen based on social-economical and demographical data from Centraal Bureau voor de Statistiek (CBS, in English: Central Statistics Office). The chosen reference areas are illustrated in Figure 3-1. The blue municipalities indicate municipalities in the risk area, the dark green indicate the first reference areas and the light green indicate the second reference areas. The yellow municipalities were determined not to be suitable for a comparison.

Figure 3-1: Risk and Reference areas

Francke & Lee (2013) state that the reference areas do differ from the risk area in a few ways. First, the number of residents in the reference areas is increasing whereas it is decreasing in the risk area. Also, the number of sold houses as a fraction of the total number of houses is lower in the risk area than in the reference areas, as well as the WOZ value of the houses. Despite these differences, Ortec Finance decided to use these reference areas.

(17)

16 characteristics of the property (Tyrvainen & Miettinen 2000). A database from the NVM was used as input for this pricing method. The data is used from 1985 to present. The type of hedonic pricing method used was the hierarchical trend model. This model should be fit for those situations in which limited number of sales in an area have occurred (Francke & Vos 2004). The goal of the model is to show the influence of property specifications, trends and small fluctuations in price, also called noise. The second method is the repeat-sales method. This model compares prices of the same property with each other each time it has been sold. In this way, it can be determined whether an increase or decrease of the value of the property has occurred. The downside of this model is that the property has to be sold at least two times before it can be used. This was the case for 50% of all sales. For this method, data from the Kadaster was used from 1993 to present. Both models were used to analyse the prices in the risk area and reference areas. This was done per quarter as this was seen best suitable for the number of sales in this timeframe. The scale used for the comparison with the reference areas is based on municipalities, the reason for this is that they find that there is too little information available to increase the zoom level. The main conclusions of the report were based on the hedonic pricing method. However, the report also includes alternative market indicators. These indicators are:

1. Number of sold properties

2. Number of NVM sales relative to number of houses 3. Number of houses put up for sale

4. Number of days a property was put on for sale

5. Number of days an unsold property was put on for sale 6. Relative difference between asking and selling price 7. Speed of sale

8. Pullback speed

(18)

17 1. Number of sold properties is lower in the risk area

2. Number of NVM sales relative to number of houses is lower in the risk area 3. Number of houses put up for sale is higher in the risk area

4. Number of days a property was put on for sale is higher in the risk area

5. Number of days an unsold property was put on for sale is higher in the risk area 6. Relative difference between asking and selling price is higher in the risk area 7. Speed of sale is lower in the risk area

8. Pullback speed is the same in the risk area and reference areas

The alternative market indicators show that the housing market in the risk area is in a poorer condition as opposed to the housing market in the reference areas even though this cannot be seen in the results from the hedonic pricing method and the repeated-sales model.

The expected goals that have been set seem to match partially the goals that can be derived from the report from Ortec Finance. The first goal states that the Ministry of Economic Affairs is likely to want to use a calculation method in the monitoring system, which minimises the depreciation of real estate value. The conclusion of the report is that there is no significantly larger depreciation of real estate value in the risk area compared to the reference areas. This by itself does not necessarily mean the report used a calculation method which decreases the depreciation of real estate value due to the winning of gas, however there is a lot of criticism from other stakeholders who state that statements based on the report from Ortec Finance are invalid and are minimising depreciation of the value of real estate (RTV Noord 2014c). Though this indicates a method could have been used, which minimises depreciation, it can’t be stated that this is an actual goal of the Ministry of Economic Affairs and thus this goal cannot be verified.

(19)

18

3.5 Verified Goals and CSF’s – NAM

The Minister of Economic Affairs requested the NAM to pay a compensation to housing owners in the risk area that have sold their property after the 24th of January 2013 (NAM 2014b). The Minister stated that the NAM is responsible for the process of determining a possible decrease of real estate value. The NAM only compensates those who live in certain municipalities, namely Appingedam, Bedum, Delfzijl, Eemsmond, Loppersum, Sloochteren, Ten Boer and Winsum. In order to determine whether there is a depreciation of the value of the property due to the winning of gas, the NAM uses a more extensive method than Ortec Finance. This method, which is being used at this moment, has recently been evaluated (NAM 2014b). According to the NAM, the persons evaluating the method have marked the current method as correct. The NAM calculates the value for each property individually and thus not for the whole region.

(20)

19 will be to pay the difference in percentage times the selling price of the property to the property owner. Before this advice however is given, the appraisers compare these results to research performed by Ortec Finance and an investigation by Calcasa. This organisation has performed an investigation about the housing market in the risk area and 10 reference areas and brings out a report each quarter to the NAM on scale of municipalities. Also, the appraisers can take into account other value influencing factors such as when the property has had a history of damage caused by earthquakes. The appraisers will send an advice about whether and how much should be refunded to a quality team. If the quality is sufficient it will be send to the seller and the NAM. If the quality team does not agree with the appraisers, the appraisers can optionally change their advice. The NAM is not required to follow this advice and can also change the advice with underpinnings to the seller. If the seller is not willing to accept the offer by the NAM, there will be a second opinion possible. This second opinion entails the reappraising of the property by three appraisers, one of which is chosen by the seller, one by the NAM and one by the chosen appraisers. Both the NAM and the seller can also reject this advice. The NAM will evaluate this method from time to time. Also, advisory reports will be evaluated randomly by the Nederlands Woning Waarde Instituut (NWWI, in English: Dutch Living Value Institute).

(21)

20 second goal is beyond the scope of this research and is therefore rejected. The third goal is valid for the NAM. The current method has a second opinion option for both the NAM and the resident.

3.6 Verified Goals and CSF’s – Province of Groningen

In order to verify the goals of the Province of Groningen, and thus those from the local authorities, an interview has been conducted with three employees of the Province of Groningen. The first interviewee was Jeroen Bakker, he is employee of program and project management. The second interviewee was Huub Hansen, he is policy officer for the community and is involved in the earthquake file of the province. The third interviewee was Werna Udding, she is involved in a project that aims to increase value of properties in the risk area.

(22)

21 The Province of Groningen is currently working on some projects related to the problem described in this research. Firstly, they are making a database from their Geographic Information System (GIS). The initial database will contain information, which allows them to check every location of the risk area in the Province of Groningen for the presence of a property, its size, the year it is built, the WOZ value and the type of property, for example an apartment. They also want to include data from the Kadaster to know whether it is a rental property or not. They are also trying to include data from the NVM and the damage claims filed by the NAM by residents in order to further complete the database. The goal of this database is to analyse the problems related to earthquakes in certain areas even better. Specifically they are interested in what the impact is of investments made to enforce properties in the risk area. They can analyse this as there is data available related to the Peak Ground Acceleration (PGA) in the risk area. This PGA data shows how strongly the earthquake shakes on a certain location and thus can give an indication of the strength of the earthquake (Saadeghvaziri et al. 2010). The initial database only contains data from the risk area, but this might be more expanded in the future. The Province of Groningen does think it will be difficult to obtain data from the NAM related to the damage claims done by residents as this might infringe with their privacy. At an aggregated level it might however be possible. Another research is related to the expected vacancy rates in the risk area and to what extent this can influence the decline in value of properties. One of the reasons for this is that it is expected that Delftzijl will reach a vacancy rate of 10% within 5 years. The Province of Groningen states it would be likely to share data with the University of Groningen, however this must be further investigated. Another research that is performed by an external party but is initiated by the province is related to investigating whether the current claims settlement process is correct. There are no results yet from this research. The Province of Groningen does have its questions with the current way the NAM deals with the claims settlement process. Jeroen Bakker mentions that as a seller you must blindly trust the NAM for compensating when a property is sold with a loss. Therefore it is difficult for sellers to put their property up for sale for a lower price. The current point at which compensation is paid can therefore be questioned.

(23)

22 layer. The Province of Groningen would benefit from a monitoring system as investigated in this research. Their goal would be to provide more information to municipalities and residents such that damage claims can be made and policies can be adjusted. As an example the Province of Groningen states that municipalities are getting more trouble getting taxes as these are paid based on a percentage of the value of a property. The responsibility of such a system would be that of the Province of Groningen.

The expected goals that were determined for the local authorities and thus for the Province of Groningen were that the system should use a calculation, which (1) maximises depreciation of real estate value, (2) takes into account emotional damage and (3) uses multiple calculation methods. The first goal can be verified for the municipalities, it can however not be verified for the Province of Groningen. Though the second expected goal might be a goal, it is not suitable to implement it in a system, which solely measures change in real estate value. It will therefore not be treated as a goal. The third goal can be verified in such that the Province of Groningen is likely to want a second opinion possibility if the outcome is low. An additional goal is found based on the interview. This goal entails that the Province of Groningen wants the system to provide information that can be used by municipalities and residents to claim damage caused by the winning of gas.

3.7 Verified Goals and CSF’s – Local Authorities: Municipalities

In order to verify the goals and CSF’s of the municipalities and thus those of the local authorities, two interviews have been conducted. The first interview was conducted with Jinko Rots, project secretary earthquakes and gas winning for the municipality of Loppersum. The second interview was conducted with Paul van den Berg, director residential and area concerns for the municipality of Eemsmond. The outcomes of these interviews were for the largest part in line with each other. Therefore, this section will first describe the outcomes of both interviews at once.

(24)

23 whether they think there is a value decline due to the winning of gas and if so, on which facts this statement is based. Both municipalities do believe there is an influence of the winning of gas on real estate value in a negative way. They base this statement on a conclusion, which is also given in the Ortec Finance report, namely that it is very difficult to sell a property in terms of duration it is for sale and prices that are paid. These are concerns the municipalities hear from both realtors and residents. Due to this, both municipalities do not agree with the main conclusion of the Ortec Finance report that there is no significant decline in value due to the winning of gas. They state that people possibly don’t even try to sell their property due to the bad market or properties are withdrawn from the market as they cannot be sold. These points are not taken into account into the main conclusion of the Ortec Finance report as those are based on sold properties. Both municipalities do think the Ortec Finance report is complete in terms of features of the property taken into account. However, the conclusions based on the data are considered to be invalid. Both municipalities do see added value in a monitoring system. They state that such a system could help them informing the residents about the consequences of the gas winning but it can also help the municipalities claim damage at the NAM as they are also experiencing difficulties in selling lots to possible residents. They state that it takes longer before lots are being sold if they are sold at all. Both municipalities indicated a monitoring system should be designed and run by an independent organisation such as the University of Groningen in order to gain the most trust. Jinko states that the municipality of Loppersum received multiple requests to lower the WOZ value because of the earthquakes that have taken place. A number of these requests were approved. These requests concerned properties that were damaged by the earthquakes and thus lost value but also concerned requests stating that property lost value due to reputational damage of the area. Both municipalities find it difficult to answer the question whether the chosen reference areas are correct. They do find the scale level of the comparison of the risk area with the reference areas of Ortec Finance too big, however, they fear that if a lower scale will be used, it would be very difficult to draw statistical significant conclusions, as there is fewer data available when lowering the scale to postal code. A way to include the intensity of the earthquake in the monitoring system according to the municipalities is the scale of Richter, the number of claims by residents but also the ground acceleration of the earthquakes.

(25)

24 multiple calculation methods. The first goal can be verified due to the fact that municipalities themselves also experience damage due to the winning of gas in terms of unsold lots they have, therefore, they are likely to want to use a calculation that maximises depreciation of real estate value and thus maximises the claim that can be made. Though the second expected goal might be a goal for a system, it is not suitable to implement it in a system, which solely measures change in real estate value and not emotional damage. It will therefore not be seen as a goal. The third goal can be verified in that the municipalities would like a second opinion possibility if the outcome is insufficient.

3.8

Verified goals and CSF’s – Residents: VEH

In order to verify the goals and CSF’s of the VEH, one interview has been conducted. The interview was conducted with Steven Wayenberg, legal policy advisor and collective interest representative in the earthquake project team at the VEH. The outcomes of the interview will be given in the following section.

The main result from the interview is that the VEH also does not have a calculation method to determine the influence of the winning of gas on real estate value. They however are convinced real estate value is influenced by the winning of gas and are thus not convinced the conclusion of the Ortec Finance report is valid. To try to show this, they hired the TU Delft to analyse the report from Ortec Finance in order to see what could have been done better. They however concluded that for the largest part, the report is correct. The main conclusion from the TU Delft was that the number of transactions that has taken place related to the selling of properties is relatively low. Due to this, it is harder to draw statistical significant conclusions. VEH also thinks that there are a lot of properties that cannot be sold in the first place and are thus not included in the hedonic pricing method used by Ortec Finance as this only takes into account sold properties. At this moment, the VEH does not have any ongoing research related to calculating the influence of gas winning on real estate value. They however did ask WOZ Specialisten, an organisation that helps lower WOZ values, to make a test case of the Groningen area in order to determine whether municipalities take into account earthquakes in making the WOZ values. They concluded that this is the case and that most municipalities thus take into account extra decrease in value due to the winning of gas.

(26)

25 residents stated their concerns about the possible lower value of their properties. Statements related to the decrease of value of properties in the risk area made by VEH are mostly based on such results and alternative market indicators such as a longer time it takes to sell a property. The VEH is also not satisfied with the current way the NAM deals with the claims settlement process related to the decrease of value. They find it strange that the settlement can only be done if the property has been sold. Instead, they propose a method in which a settlement is guaranteed for all the residents, thus giving them more security. The VEH does think the included property characteristics in the Ortec Finance report are sufficient for determining the value of an individual property. They however cannot say whether the chosen reference areas are correct. They also find it hard to determine whether the chosen scale for the comparison in the Ortec Finance report is sufficient, they however do state that there could be differences within municipalities and that it would thus be more logical to compare postal codes with each other instead of municipalities. Just the scale of Richter according to the VEH should not determine the impact of earthquakes in a monitoring system. They state that more than just the intensity of an earthquake could affect the value of properties. As an example they state that it could be that a certain municipality has a bad reputation thus possibly decreasing the value of real estate even more. The VEH is very interested in the proposed monitoring system. They state it could really help their members in the claims settlement process and that it can give them a more secure feeling that a grounded claim can be made if a member wants to sell its property. The GBB also finds that such a monitoring system should be developed and maintained by an independent party such as the University of Groningen. The involvement of the NAM should be kept to a minimum, as residents do not have faith in them at this moment.

(27)

26 security to housing owners in the risk area in a sense that if they want to move, they do not have to face the damages due to a lower value of their property.

3.9 Verified Goals and CSF’s – GBB

In order to verify the goals and CSF’s of the GBB, Lambert de Bont was interviewed. He is board member of the GBB and is also participating in the dialogue table. The interview was conducted in person at the UMCG.

(28)

27 and the risk area might be comparable, however there are more factors that are not included such as the atmosphere in the different areas. The GBB doubts whether the reports made by Ortec Finance are truly made without taking into account the interests of the Ministry of Economic Affairs. They state that it does look that the report of Ortec Finance is supporting the policy of the Ministry of Economic Affairs. What Ortec Finance should do instead is to also include damaged properties that have not been sold in their analysis. Though Ortec Finance did make such a report, they only included heavily damaged properties in its calculations and they did not include moderately damaged properties. The GBB finds it difficult to say whether the chosen aggregation level of Ortec Finance, which is municipalities, is the correct one. They do think it is necessary to be able to draw statistical significantly based conclusions and thus that there must be enough data available to analyse. In order to measure the impact of earthquakes, the GBB thinks PGA data must be used. This PGA data only measures horizontal speed of an earthquake, the GBB thinks that a variable for the vertical speed must also be included in order to fully measure the impact of an earthquake. The GBB thinks a monitoring system for real estate value in the risk area should be the responsibility of an independent party and thus not the NAM. They think there is currently no organisation that could be responsible for such a task and that such an organisation thus must be made. To add to this, the GBB states that even the realtors aren’t as independent as one might think. The GBB states that they are willing to do a lot in order to work for the NAM. Based on this interview, it can be stated that the VEH and the GBB have a lot in common when looking at goals. They both stand for the interests of their members, which both are property owners in the risk area. Therefore, the goals and their related CSF’s are determined to be the same for the VEH and the GBB.

3.10 Verified Goals and CSF’s – General Goals and CSF’s

(29)

28 Table 3.2: Summary of Verified Goals and CSF's

Stakeholder Goals CSF’s - Functional CSF’s – Non-Functional

Ministry of Economic Affairs

 Uses a calculation that uses multiple calculation methods (second opinion)

 Increase knowledge of effects of winning of gas on real estate value

The system shall use multiple

calculation methods and shall deliver knowledge increasing information

The system shall be flexible enough to incorporate multiple calculation methods

NAM  Uses a calculation that uses multiple calculation methods (second opinion)

The system shall use multiple calculation methods

Province of Groningen

 Uses a calculation that uses multiple calculation methods (second opinion)

 Should output usable info for applying damage claims

calculation methods and shall output a grounded, validated value for

depreciation

Municipalities  Uses a calculation that:

o Maximises depreciation

o Uses multiple calculation methods (second opinion)

calculation methods and shall maximise the calculated depreciation

Residents - VEH

 Uses a calculation that:

o Uses multiple calculation method (second opinion)

 Should bring security to housing owners in the risk area

calculation methods and shall maximise the calculated depreciation. It shall output a grounded, validated value for depreciation

The system shall be flexible enough to incorporate multiple calculation methods. It shall be accepted by all stakeholders thus bringing security to housing owners

Residents - GBB

 Uses a calculation that:

o Uses multiple calculation method (second opinion)

 Should bring security to housing owners in the risk area

calculation methods and shall maximise the calculated depreciation. It shall output a grounded, validated value for depreciation

The system shall be flexible enough to incorporate multiple calculation methods. It shall be accepted by all stakeholders thus bringing security to housing owners

All Have a system that:

 Outputs usable data

 Outputs complete data

 Is reliable

 Needs low maintenance

 Is flexible to changing context

The system shall bundle data from multiple sources to output usable, complete data for its users.

(30)

29

4 DIAGNOSIS AND ANALYSIS

This chapter will start by shortly describing the main outcomes of the conducted interviews. After that, the problems related to resolving each CSF will be described. There appears to be no order-dependency. All CSF’s must be met for the system to properly function.

4.1 Main outcomes

One of the most interesting outcomes from the interviews is that besides the NAM and Ministry of Economic Affairs, not a single stakeholder has a method to calculate the difference in real estate value due to the winning of gas. Also, those stakeholders only have limited criticism on how the methods used by the Ministry of Economic Affairs and the NAM can be improved. Besides the outcomes related to the goals, there are somec other outcomes from the interviews that need to be mentioned. All interviewees except the NAM agreed that a method that could monitor real estate value should not be the responsibility of the NAM, it should be done by an independent party, for example the Centrum Veilig Wonen (CVW, in English: Centre Safe Living). The CVW is currently responsible for handling claims made by residents concerning physical damage to properties. Also, the current method used by the NAM, which also uses the report made by Ortec Finance, is not seen as an appropriate method. Some points of improvement are given by the stakeholders, which are further described in chapter 5.

4.2 Functional CSF’s

1. The system shall use multiple calculation methods – This CSF applies for all stakeholders, it must therefore be possible for the monitoring system to use multiple calculation methods or in other words, it must have a second opinion possibility. The difficulty in resolving this CSF is to find at least two calculation methods that will be acceptable to all stakeholders. An answer to this question might lie in the report made by Ortec Finance. This report uses two calculation methods, namely the hedonic pricing method and the repeat-sales method. Though most stakeholders do not support the conclusion from the Ortec Finance report, the method used by Ortec Finance is seen as a valid method by all stakeholders except for the GBB. The GBB however does not have an alternative at hand and it could therefore be possible to adjust the method used by Ortec Finance in such a way that all stakeholders will accept it.

(31)

30 initiated by the Ministry of Economic Affairs. In order to attain knowledge increasing information, it must be clear what is defined as knowledge increasing information. In this research, this can be defined as information that shows the change in value of real estate due to the winning of gas. Therefore, the system must show the difference between the actual value of a property and the value of a property if there were no earthquakes. Initially this would have been done by comparing different calculation methods of the stakeholders. However, only the NAM and Ministry of Economic Affairs have a calculation method thus making this impossible. What can be done is to incorporate the criticism of the other stakeholders related to the current calculation method in a proposed method. This will be shown in chapter 5.

3. The system shall maximise the calculated depreciation – The Province of Groningen, municipalities, VEH and GBB want a calculation method that maximises depreciation. The difficulty in resolving this CSF is that there must be found a balance between minimising and maximising the calculated depreciation such that all stakeholders accept it. All stakeholders must therefore validate the proposed method before it will be usable.

4. The system must produce a grounded, validated value for depreciation – This CSF is extracted from a goal of the Province of Groningen, namely that the output should be usable for damage claims. In order to be usable, the system must output a depreciation value that will be accepted by all stakeholders and thus a validated calculation method must be used in the monitoring system.

5. The system shall bundle data from multiple sources to output usable, complete data – This CSF is extracted from the goals all stakeholders will have, namely that the system produces usable and complete data. In order for the data to be usable and complete, it must be clear what is defined as usable and complete. In this case, it means that the system specifies a value of a property that will be accepted by all stakeholders and therefore is based on data from multiple sources in order to ensure a correct calculation.

4.3 Non-functional CSF’s

(32)

31 possible to process data from multiple sources. Therefore, multiple organisations aside from the stakeholders must cooperate in order for the system to fully function.

2. The system shall be accepted by all stakeholders- A must for all stakeholders is that the system and its outcomes are acceptable. If this is not the case, the whole system will be useless as it will not be possible to claim any lost real estate value. The hardest part in this is that all stakeholders must accept the calculation method used, also see CSF number 3 of the functional CSF’s.

3. The system shall be reliable, low to maintain and shall be flexible to a changing context – A few non-functional CSF’s can be identified for all stakeholders, namely that the system shall be reliable in terms of output but also in uptime. It is therefore necessary to ensure the output is constantly monitored for correctness and that the system shall be run by a party who has backup facilities in case of a system failure. Besides that, the context of the problem is constantly changing, as more parties get involved in the problems related to the winning of gas. It must therefore be possible for the system to adapt to this changing context. Therefore, it must be possible to incorporate new sources of data if this is necessary. Also, if new calculation methods are available, it must be possible to incorporate them. The difficulty in this is that the system should be easily adjustable.

4.4 Implication of results on design solution

(33)

32

5 DESIGN SOLUTION

At this point in time, there is only one alternative available, namely the calculation method that the NAM uses to accept or reject damage claims for individual properties, this calculation method includes the method made by Ortec Finance. In order to build a proposed model, the method used by the NAM will be improved based on comments by all stakeholders. Therefore, the current model of the NAM will be firstly shown in BPMN such that the improvements based on the comments of all stakeholders are clearly visible in the proposed model. This proposed model can be used as the base for a system that can monitor the influence of gas winning on real estate value.

5.1 Current NAM model

The current NAM model as described in section 3.5 is the base for the suggested improvements. The overview of the NAM model has been drawn in BPMN and can be found in Appendix E. The input coding and output decoding steps are shown in this model, the input transformation, which is in this case, the appraisal of the property is shown in Figure 5-1.

5.2 Proposed model

During the interviews, the interviewees mentioned several points of criticism about the current NAM method for claims but also for the Ortec Finance report, which is also used by the NAM in the claims settlement process. Each of these points will firstly be described after which a proposed model can be derived from them.

(34)

33 1. Point at which claim can be made

One major point of criticism that has been given related to the current NAM method is that a claim can only be made once a property has been sold. The first problem that arises in this is that individuals, who cannot sell their property at all, are excluded from the claims settlement process. Secondly, it is uncertain for property owners what amount will be paid once the property is sold. Therefore it is difficult for property owners to set a bottom price for their property to for example ensure no outstanding debt is present when selling their property. One way to solve this problem is to move the point at which a claim can be made to the point at which an individual puts up its property for sale. One thing that needs to be taken into account when moving this moment is that it should not be possible for the seller and buyer of the property to make agreements about the selling price and thereby swindle the NAM.

2. Let customers choose realtor in first offer

Currently, Arcadis chooses both realtors the first time a property will be appraised for the claims settlement process. In the second opinion possibility it is however possible for the property owner to choose its own realtor. It seems fairer when this opportunity is also given to the property owner in the first round it is appraised.

3. Non-sold properties are not taken into account

(35)

34 could show this, are the number of sold properties and the number of houses put up for sale. A third factor that has to be included is the total amount of properties in the risk and reference areas. Once this data is available, a ratio measuring the problem can be made available. Based on the outcome of the current alternative market indicators it is highly likely that this ratio will show that more properties in the risk area cannot be sold as opposed to properties in the reference areas (Francke & Lee 2014b). This ratio however cannot directly be used when determining the amount of money that has to be paid to the person filing a claim, as this ratio is not expressed in monetary terms. Therefore it must be investigated how this ratio can be used to increase or decrease the amount of money that will be paid to the person filing for the claim. One way this can be done is by letting this ratio influence the data related to the selling price trend in the risk area. If there are a lot of unsold houses in the risk area, one could lower the selling price trend in the risk area such that there will be a larger difference in the selling price trend in the risk area opposed to the reference areas. This way, the amount of money that will be paid to persons filing for a claim will become higher, as the amount of money that will be paid is calculated by multiplying the selling price trend difference in percentage with the selling price of the property.

An alternative or extension to the solution described above to solve this third problem is by looking at the number of vacant houses and/or number of foreclosures. According to Turnbull & Zahirovic-Herbert (2011), vacant properties stand less strong on the housing market and are likely to be on sale for a longer period of time. They describe to what extent vacancy affects the selling price in monetary terms. Because properties in the risk area are longer for sale, people are likely to move to another property without their first property being sold thus leaving it vacant. The costs of this can be calculated according to the method described by Turnbull & Zahirovic-Herbert (2011) and could possibly be taken into account in the monitoring system.

(36)

(37)

36 Figure 5-3: Proposed Model

Changed claim start

(38)

37

6 CONCLUSION AND DISCUSSION

A Dutch version of this chapter can be found in Appendix F. This research has shown the main goals of all stakeholders for a system that can monitor the influence of the winning of gas on real estate value and has shown improvements on the current claims settlement process of the NAM. A shared goal for all stakeholders is that a monitoring system should have a second opinion possibility when the outcomes are not accepted. Furthermore, the output should be usable for those who want to file a claim against the NAM regarding a loss in value due to the winning of gas. Except for the NAM and the Ministry of Economic Affairs, the current claims settlement process is seen as incorrect as the paid claims are not high in monetary terms. These stakeholders are also of opinion that the output of the monitoring system must maximise depreciation of properties due to the winning of gas. All stakeholders except for the NAM agree that such a monitoring system must be the responsibility of an independent party such as the CVW. Interestingly, the NAM is negatively towards a monitoring system. They state that it is very difficult to calculate such a thing and that it can be harmful to the housing market if a third party does this type of calculation. These goals were also mainly expected. It is likely they can be dealt with when building a monitoring system.

To be able to build a monitoring system, the stakeholders were also asked their opinion about the current way the NAM deals with the claims settlement process and how Ortec Finance calculates the influence of gas winning on the housing market. The first main conclusion that can be drawn from this analysis is that except for the Ministry of Economic Affairs and the NAM, no stakeholders have a method to calculate the influence of the winning of gas on real estate value. This is a surprising result as a lot of criticism is given by all stakeholders to the current way of calculating the influence of the winning of gas on real estate value. Except for the NAM and Ministry of Economic Affairs, all stakeholders agree that the current method used by both the NAM and the Ministry of Economic Affairs is insufficient. They however have limited feedback on how to improve these methods. This is mainly because of a lack of capacity and knowledge of the stakeholders. The feedback that has been given is incorporated in the proposed model.

(39)

38 moved to the moment at which a property is being put up for sale. When doing this, the property owners will have more security when selling their property. The NAM is afraid of fraud against them by the seller and buyer of a property as the claim is being paid as a percentage of the selling price. This problem can be solved by paying the claim as a percentage over a predetermined market value or when paying a fixed amount of money. Secondly, the current claims settlement process only lets property owners choose their realtor for the appraisal in the second opinion option. Instead, it should be possible for property owners to choose their realtor for the appraisal process the first time the property is being appraised. The NAM states they do this because it is seen as normal in such claims settlement processes. This however cannot be seen as a reason to exclude the property owner from choosing their own realtor to appraise their property.

The last point of feedback can be applied to both the NAM and the way Ortec Finance calculates the real estate value. The main point of criticism on both methods is that it only takes into account sold properties. In the case of Ortec Finance, this means that only sold properties in the risk area are compared to sold properties in the reference areas. Therefore, when an individual cannot sell its property, this is not taken into account. The same yields for the NAM method. This method firstly compares the sold property to similar sold properties in the same area and compares the trend of the selling prices over a period of 5 years with the trend of selling prices in reference areas over that period. The criticism is that it can be possible that in the area of the property itself, the trend of selling prices is less disturbed because for example only properties in very good shape are sold, therefore camouflaging the actual price trend in the area. This problem can possibly adding an extra factor to the calculation method, namely by comparing the ratio between properties that are put up for sale and properties that are actually sold in the risk area and the reference areas. It must be further analysed as to how this ratio can influence the claim. This may cause difficulties, as such a ratio is not expressed in monetary terms. An alternative or extension to the solution described above to solve this third problem is by looking at the number of vacant houses and/or number of foreclosures. The advantage of this solution is that it can be described in monetary terms based on the work of Turnbull & Zahirovic-Herbert (2011).

(40)

39 be interviewed for each stakeholder. Another limitation is that the proposed model in this research has not been validated by all stakeholders. It could therefore be an insufficient improvement. Future research can solve this by validating the model.

(41)

40

7 REFFERENCES

Balsters, H., 2013a. Data & Process Integration: Deriving Business Data from Business Processes, with an application to Data Management in Energy, Groningen, The Netherlands: University of Groningen.

Balsters, H., 2013b. Mapping BPMN process models to data models in ORM, Groningen, The Netherlands.

Balsters, H., 2014a. A System for extracting Data Models from basic Business Process Models, Groningen, The Netherlands.

Balsters, H., 2014b. Design Methods : Building utility-driven Artifacts Part 2, Groningen, The Netherlands: University of Groningen.

Benjamin, J.D. et al., 2002. Technology and Realtor Income. Journal of Real Estate Finance and Economics, 25(1), pp.51–65.

Berkhout, T.M. & Hordijk, A.C., 2010. Marktwaarde als waarderingsgrondslag, Delft, The Netherlands.

Centraal Bureau voor de Statistiek, 2014. Corporate Informatie. cbs.nl. Available at: http://www.cbs.nl/nl-NL/menu/organisatie/corporate-informatie/default.htm. Clapp, J.M. & Giaccotto, C., 1992. Estimating Price Indices for Residential Property : A

Comparison of Repeat Sales and Assessed Value Methods. Journal of American Statistical Association, 87(418), pp.300–306.

Dagblad van het Noorden, 2014a. Stad zucht diep onder reeks aardbevingen. Dagblad van het Noorden, p.1.

Dagblad van het Noorden, 2014b. Strijd om bevingsgeld. Dagblad van het Noorden, p.1. Available at:

http://www.dvhn.nl/nieuws/groningen/strijd-om-bevingsgeld-11885773.html.

De Kam, G.R.W. & Elhorst, J.P., 2014. Inzet RUG voor onderzoek naar impact aardbevingen op wonen en welbevinden, Groningen.

Francke, M.K. & Lee, K.M., 2013. De waardeontwikkeling op de woningmarkt in

aardbevingsgevoelige gebieden rond het Groningenveld, Rotterdam, The Netherlands. Francke, M.K. & Lee, K.M., 2014a. De ontwikkelingen op de woningmarkt rond het

Groningenveld: actualisatie 4e kwartaal 2013, Rotterdam.