MIS 9003 – Prof. Min-Seok Pang

Xi Wu

Hello, everyone. I am a phD student in Management Information Systems concentration. I received a B.S. in Automation from Central South University and M.S in Pattern Recognition & Intelligent Systems from Huazhong University of Science & Technology. Prior to joining Fox business school, I was working as a Research Assistant in the field of Management Information System in Chinese University of Hong Kong business school and got a MSc degree in Business Analytics from Chinese University of Hong Kong. My research interests lie in economics of IT. Some specific topics include e-commerce strategy, online marketing and enterprise software. I am open to discuss any interesting topics in IS field.

Week 13 Reading Summary – Pang and Tanriverdi (2017) – Xi Wu

Pang and Tanriverdi (2017) Security Breaches in the U.S. Federal Government

There has been limited research on the mitigation mechanisms of security vulnerabilities with actual breach incident data at the organization level. This paper studies the effectiveness of three organizational IT risks mitigation mechanisms: modernization of legacy IT systems, institution of effective IT GRC, and migration of legacy IT system to the cloud. The hypotheses are developed based on criminology theories such as rational choice theory and crime opportunity theory. (1) Federal agencies that spend higher percentages of their IT budgets on the maintenance of legacy IT systems are likely to experience more security incidents than ones that spend higher percentages of their IT budgets on IT modernization and new IT development; (2a) IT GRC effectiveness reduces security incidents in federal agencies; (2b) IT GRC effectiveness substitutes IT modernization in reducing security incidents; (3) Federal agencies that migrate their IT systems more to the cloud are likely to experience fewer security incidents.

The unit of analysis is a U.S. federal agency. Data on security incidents in the federal agencies is obtained by FIMSA report. Security breaches to the federal agencies in 2005-2016 is obtained from an independent source PRC. Data on the IT investment profiles of federal agencies is collected from the Federal IT Dashboard.

The study finds that (1) A 1%-point increase in the share of IT modernization in the IT budget is associated with a 5% decrease in security incidents. (2) the institution of effective IT GRC mechanisms significantly reduces the security incidents. (3) a negative interaction effect between IT modernization and IT GRC. The findings complement the extant IS security literature on the technical mitigating mechanisms by assessing the effectiveness of more managerially actionable mechanisms.

Week 12 Reading Summary – Menon and Kohli (2013) – Xi Wu

Menon, N. M., & Kohli, R. (2013). Blunting damocles’ sword: A longitudinal model of healthcare it impact on malpractice insurance premium and quality of patient care. Information Systems Research, 24(4), 918–932.

Few studies have considered its implication on product or service quality, this study fills the gap by investigate the impact of healthcare IT (HIT) expenditure on the malpractice insurance premium (MIP) and the moderating effect of HIT expenditure on the relationship between MIP and patient care quality. Based on the prior literature of IT value and risk management, three hypotheses are proposed. 1) Past HIT expenditure is positively associated with current quality of patient care. 2) Past HIT expenditure is negatively associated with MIP. 3) Past MIP is positively associated with quality of patient care. 4) The relationship between past MIP and quality of patient care is enhanced (positively moderated) by past HIT expenditure.

There are multiple data sources in this study. Hospital information is from the Washington State Department of Health. Patient care outcomes are obtained from a data services and consulting company. Three regression models are applied. To control for the likely persistence in organizational decisions and actions from year to year, lag of dependent variables is added. First-differencing method is applied to address the fixed effect factors. Some instrument variables are used to overcome the endogeneity problem. Results support all the hypothesis but H4. The moderating impact of HIT on quality of patient care through its impact on the link from past MIP to quality of patient care is negative. The findings offer opportunities for future research.

This study contributes to understanding the expectation of IT benefits and its effect on an organization. It also informs decision makers in risk, quality management, and the IT function to engage in joint risk mitigation decisions to achieve desired organizational goals.

Week 11 – Reading Summary – Tambe and Hitt (2014) – Xi Wu

Tambe, P., & Hitt, L. M. (2014). Job Hopping, Information Technology Spillovers, and Productivity Growth. Management Science, 60(2), 338–355.

Like earlier general-purpose technology such as electricity or the steam engine, researchers argued that information technology (IT) investments generate productivity “spillovers” among firms, and the movement of IT workers among firms is believed to be an important mechanism by which IT-related innovations diffuse throughout the economy. In this paper, the authors use employee microdata obtained from an online resume database to test the hypothesis that firms benefit from the IT investment of other firms because the flow of specialized technical know-how among organizations facilitates the implementation of new IT innovations.

Based on the literature on the impact of R&D spillovers and IT investment on productivity, this study modeled the knowledge available to the focal firm as the weighted sum of the knowledge of other firms in the sample, and the transfer of IT-related knowledge occurs through the mobility of workers. Two IT investment measures are applied: IT capital stock-based measure and IT labor-based measure. The external IT pool measure is computed as the IT intensity of other firms. OLS and Fixed Effect model are used to support the hypothesis of productivity spillover effect through the IT labor flow. This study also compares the benefits through IT labor flow with that through the IT investments of geographically proximate firms, the estimators indicate that regional spillovers appear to be driven by IT labor flows, and IT labor flows appear to be an important source of spillovers even outside a fixed region.

This study is the first to analyze how IT labor flows drive IT spillovers and, to investigate this issue using microdata on labor mobility.  It suggests that a substantial amount of variation in IT returns can be explained by productivity spillovers generated by IT labor flows.


Week 9 Reading Summary – Pang et al. (2016, Mgt Sci) – Xi Wu

Pang, M.-S., Tafti, A., & Krishnan, M. S. (2016). Do CIO IT Budgets Explain Bigger or Smaller Governments? Theory and Evidence from U.S. State Governments. Management Science, 62(4), 1020–1041.

There is a fundamental economic debate among the public, the media, and academia as to what is the proper boundary and role of a government in the contemporary economy. Technological development as one of the key factors behind government growth, should be paid more attention. This study is interested in how information technology (IT) investments made by governments affect the size of government spending. More specifically, this paper examines the relationship between IT budgets of a state chief information officer (CIO) and state government size.

Two opposing hypotheses are proposed. One is that IT investments make state administration more efficient, productive, and transparent, leading to a reduction in state expenditures. The other one is that state governments can utilize digital technologies in initiating a new range of public services that would not have been offered without IT and having a greater control over functions that external organizations would otherwise handle, thus increasing government spending.

Using a variety of data on IT spending and state government expenditures, this study finds that greater IT investments made by a state chief information officer (CIO) are associated with lower state government spending. It is estimated that on average, a $1 increase in state CIO budgets is associated with a reduction of as much as $3.49 in state overall expenditures. Dynamic panel-data model is applied the empirical analysis. A series of robustness checks are implemented to provide further evidences and support the hypothesis.

This work contributes to the IS literature by expanding the boundary of IS research on IT and organizational size to the public sector organizations, to which the IS research has not paid as much attention as it has to the business setting.

Week 8 Reading Summary – Tiwana and Kim (2015) – Xi Wu

Tiwana, A., & Kim, S. K. (2015). Discriminating IT Governance. Information Systems Research, 26(4), 656–674.

Strategic IT agility is one important weapon for firms that using IT to consistently create a series of temporary advantages, introducing a new one before rivals could even finish copying the last one. This study asks the question: why are some firms more adept at using IT in their pursuit of strategic opportunities? The belief is that the secret for exploiting IT for strategic agility is how it is governed, i.e., which department makes what IT decisions. This study addresses the research question: How does the interplay between firm’s IT governance choices and departmental peripheral knowledge influence IT strategic agility.

The theoretic foundation is the JM theory that decision rights must be collocated with the knowledge needed to make those decisions. This study develops the idea that firms’ IT governance choices foster IT strategic agility only when their alignment with departments’ peripheral knowledge is discriminating—discriminating in that only a specific department’s peripheral knowledge induces agility for a specific class (apps or infrastructure) of IT decisions; which department has peripheral knowledge must be aligned with which department makes what IT decisions.

Matched-pair data from 105 U.S. firms are collected. All principal constructs using reflective multi-item Likert scales are measured using the firm’s IT function as the unit of analysis. To account for the endogeneity in firms’ IT governance choices, several instrumental variables are used. The two hypotheses about IT app governance and IT infrastructure governance are theoretically support by a three-step hierarchical weighted least squares (WLS) model. This study contributes theoretically to discriminating alignment and governance-contingent value of peripheral knowledge in IT governance literature.

Week 7 – Reading Summary – Gopal and Koka (2012) – Xi Wu

Gopal, A., & Koka, B. R. (2012). The Asymmetric Benefits of Relational Flexibility: Evidence from software development outsourcing. MIS Quartely, 36(2), 553–576.

Situating the vendor-client relationship in the center of a series of economic and social interactions between organizations. As exchange hazards increase in the relationship, formal contracts are limited in their ability to rein in opportunism due to their incompleteness. Thus, relational governance plays a critical role in the relationship. There are two gaps in current views, one is assumption that these enacted relational mechanisms provide symmetric benefits to all parties. The other gap is, how is value from relational governance captured for different parties in inter-organizational economic activities.

This paper argues that the enacted observation of relational flexibility is driven by perceptions of exchange hazards. This paper proposes that the benefits accruing from relational flexibility are asymmetric and depend on how the exchange risks are apportioned by the formal contract. In addition, it hypothesizes that these benefits manifest on the performance dimensions that are of importance to the risk-exposed partner. These hypotheses are tested on 105 software projects completed by a software outsourcing vendor for multiple clients.

The results show that relational flexibility positively affects profitability in only fixed price contracts, where the vendor faces greater risk, while positively affecting quality only in time and materials contracts, where the client is at greater risk. In contrast, the impact of relational flexibility on quality is higher in T&M contracts where the client bears the greater portion of risk. The findings highlight the need to establish risk exposure first, and then examine the effects of flexibility on performance contingent on risk exposure. They also highlight the implications of relational governance for the performance dimensions of interest. In addition to the theoretical contributions, this paper also makes a methodological contribution by using a triangulating approach to examine the endogenous interactions of relational flexibility and contract type on project outcomes.

Week 6 – What makes a good teacher?

(S – Students  P – Professor)

S: I do not have one favorite teacher inspiring me, but I like some aspects of different teachers. For example, they care about the students, not just the number of scores. In my class, I try to remember students’ names. It’s hard, but it is something good to try. It shows that you care about them.

S: My favorite teacher is my advisor when I was a master student. When I started research, I knew little about it. My advisor gave me a lot of hands-on advices such as how to collect data, how to write a paper. She was busy, but she still spent a lot of time on student. Besides, she is very passionate about her research. She is a role model for me.

S: I want to talk about my Chinese verbal teacher in my 7th grade. At that time, I was very shy and not good at speak in front of people. The teacher made a lot of try to encourage me. For example, he asked us to describe a latest news in front of class every week. He tried to open my mind and speak in public. He changed my life.

P: A good teacher is someone inspired you. He or she cares about students, gives an opportunity to change the student’s life. It is hard to say which aspect is the most important, there are different aspects matter for students in different stages.

S: My favorite teacher is a mathematics teacher who helped me in the college entrance exam. With his help, I became strong in Calculus at that time.

S: I think for a teacher, it is hard to inspire every student. But there are a lot of little things you could do. You need to be organized, preparing for the class, doing a good time management, sending students emails to remind them of homework.

P: I know there are professors who have a class at 9:00. They will arrive at 9 and start to prepare for the slides, then the exact time when the class begins will be 9:05. Do not do that. You should arrive earlier for preparation work. This is a small thing, but it’s important. You need to show that you are ready, well prepared and professional.

S: I like the teachers who do not just teach based on the lecture notes or textbooks. They will extend from the text book and teach something else within the current topic. And a good teacher should be encouraging, always inspire students. They should be approachable, not only in class, but also by email, office hours though they might be busy.

P: The teacher need to be a good guy. You could not be a different person in class through the whole semester. If a teacher is not a good person, he or she doesn’t care about students. Students will know that.

S: And a teacher should think the course is important and put effort in it. If the teacher doesn’t think the course is important for them, the students would not think it’s important.

P: This is what I am doing for this class. Teaching a doctoral seminar means I need to spend one day to reading all the papers which we will discuss each week. Compared to teaching undergraduate or MBA courses, teaching doctoral seminar means more effort.

S: One teacher I like often thought of some games or tools to help us understand the difficult concepts. He was well prepared, and students will know that he put many efforts in it.

P: You mean that the class is not just about lectures, right? Here is something very important. The more you talk, the lower your teaching evaluation would be. A good teacher should interact with students, provide many different ways for students to understand. There is a study about different ways for learning: Reading textbook, listening to teachers, talking as a group, watching videos, practicing and teaching others. Reading and listening are not the most effective ways because they are passive learning methods. Teaching is a very good way. I am practicing this method right now. Each week I asked you to summarize the paper and explain it to others, it is a good way to understand the paper.  Something fuzzy might become clear when you explain to others. And I always asked you to tell us what you think in your mind. Something will become concrete when you organize it in your mind.

S: Sometimes a big name in academia will inspire students.

P: I think it is very important to inspire students, but it is difficult.

S: Doctoral students are motivated, but for undergraduate students, some of them do not care about the classes. They even do not want to make a cheat sheet for exams, it seems that they do not care about this degree.

P: This is the 2nd point I want to emphasize. It is very important to convince your students why this course is very important. Especially for MIS field, although students are forced to take core IS course, they do not think it is necessary for business students. I spent a lot of time to explain why we need to care about technology and this IS course. For those which are not core course, students choose them because they have enough motivation. But for core course, the teacher should try to convince them.

Another key ability a good teacher should have is to think from the student’s perspective. When you are designing slides and homework, think about effectiveness from a student’s perspective. Are they going to enjoy this? It is not saying that just to make students happy. But we need to make the course durable, enjoyable and student could learn something from it. One example is about deadline. If you set the homework deadline to Friday, there is always students asking whether they could have an extension.  So you could specify the penalty, for example 10% deduction after Friday submission. Things could happen to students so there should be some flexibilities.

Here is something I found useful when I am teaching. I will give unofficial grades in the middle of the semester. Because students are anxious when they are waiting for the final grading. If I tell them the unofficial grades of B- in the middle of the semester, they would be not that surprise when they find their final grading is B because they already know their performance.

S: I think it is good if you give students feedbacks. I will tell students their performance in the previous sections. If they don’t do well, they still have a chance to fix it. In addition, I will design the course to make sure it includes different ways for grading, test, group study, project, etc.

P: Business school usually emphasize active learning. They care about students’ participation, but sometimes they just kill introverts. There is an article online “When Schools Overlook Introverts”. When I was in University of Michigan, I was required to take MBA courses in Marketing and Finance. It was a nightmare. Look back the experience, I knew that someone just hate participation in class. So, I use in-class participation and online participation when I am teaching. Students love that. The principle behind this is, understand what the students think.

S: That’s what I did in class. If someone likes answering questions, let them do. If someone don’t like it, don’t force them.

P: It is true that some students will dominate the class. Sometimes I will wait and ask whether someone else would like to answer the questions. Teaching is important, you have to put effort, try to get high teaching evaluations, and develop your own best practice. When you just at the beginning, you need to take a lot of effort, but it will get easier after you figure out all of these. There is a resource which would be helpful –  Temple University Center for the Advancement of Teaching. There are several programs and events you could take. When you are doing a TA, you could learn a lot by actively involved in the class. There are two senior students sitting in my class, we will review the class together after each class. What went well, what went wrong. They could learn a lot from it.

S: I have a question about how to arrange the time for your teaching work. I think most of the research universities require teachers to teach three each year.

P: Yes. There is a tip when you are choosing the university if you have options. Life in a university with big undergraduate program will be easier. Because the same course will be opened multiple times, then you just need to prepare for one course material but teach them several times. It is good for both teaching and research.

Week 6 Reading Summary – Xue et al. (2011) – Xi Wu

Xue, L., Ray, G., & Gu, B. (2011). Environmental Uncertainty and IT Infrastructure Governance: A Curvilinear Relationship. Information Systems Research, 22(2), 389–399.

Extant research predicts that firms tend to decentralize IT governance in more uncertain environments. But empirical work in this area presents mixed result. To develop a deeper understanding of the relationship, this paper investigates the issue by studying the relationship between environmental uncertainty and IT infrastructure governance in a sample of business units from Fortune 1000 companies.

Centralization of IT infrastructure provides the benefits of economies of scale as the same IT platform can be leveraged across different business units, thereby reducing the unit cost of IT infrastructure for each business unit. But when uncertainty increases, the benefits of responsiveness provided by more decentralized IT infrastructure governance are likely to outweigh the benefits of centralization. However, decentralization of IT infrastructure governance raises issues of control. When uncertainty increases from the intermediate to high level, firms may switch from decentralization to recentralizing IT infrastructure governance.

The key proposition is that the relationship between environmental uncertainty and decentralization in IT infrastructure governance is best characterized as a curvilinear relationship. Moreover, the study proposes that business unrelatedness between business units and their headquarters moderates. The dataset is obtained from the CI, Compustat database, SDC Platinum Mergers and Acquisitions databases. Binary logistic regression model is applied to test the hypotheses. The result support all the hypotheses.

The key theoretical implication of this study is that the relationship between environmental uncertainty and IT infrastructure governance is likely to be more complex than that suggested by the prior literature. This paper also provides practical implications to managers when they make IT infrastructure governance decisions.

Week 5 Reading Summary – Kang et al. (2017) -Xi Wu

Kang, K., Hahn, J., & De, P. (2017). Learning Effects of Domain, Technology, and Customer Knowledge in Information Systems Development : An Empirical Study. Information Systems Research, 28(4), 797–811

Investigation of ISD performance by identifying the factors associated with its increase (or decrease) has recently been the focus of much attention in the information systems literature. Some literature proposes that ISD performance can be improved by taking advantage of learning curve (or experience curve) effects. This study examines how ISD project teams learn and under what conditions the learning effects are stronger or weaker. The authors conceptualize ISD as the application of various elements of domain, technology, and customer knowledge, and propose that the units of the same experience should be these knowledge elements.

The research questions are: (1) Do ISD project teams benefit from prior experiences in domain, technology, and customers in increasing project performance? (2) whether and how experiences in the three types of knowledge relate to one another. (3) How does ISD project complexity moderate the learning effects of these three types of experience?

From a dataset collected from a prominent global IT services company, this study finds that ISD project teams’ experience in prior projects translates into performance gains for the current ISD project when the prior and current projects share the same domain, technology, or customer knowledge elements. In addition, learning effects of domain, technology, and customer knowledge are substitutive for one another and that these learning effects become stronger or weaker depending on the extent of ISD projects’ team and task complexities.

The study makes significant contributions to the ISD literature on learning effects and the roles of domain, technology, customer knowledge, and project complexity, as well as to the general organizational learning literature. It also provides important managerial insights into practical concerns such as project staffing and knowledge acquisition for ISD organizations.



Week 4 Reading Summary – Chi et al.(2010) – Xi Wu

Chi et al. (2010). Information technology, network structure, and competitive action. Information Systems Research, 21(3), 543-570.

Researchers in competitive dynamics have demonstrated that firms that carry out intense, complex, and heterogeneous competitive actions exhibit better performance. However, there is a need to understand factors that enable firms to undertake competitive actions. This study focuses on the sparse-versus-dense network structure of inter-organizational networks and aim to theoretically and empirically investigate how these contrasting types of network structure interact with IT to influence firm competitive behavior. They argue that firms could benefit from both types of network structure, but the extent to which they exploit their network positions is contingent upon their use of IT.

This paper extends the AMC framework to the action repertoire level of analysis by examining how a firm’s network structure interacts with its use of IT system to jointly influence the firm’s competitive behavior through increasing the firm’s awareness of opportunities for developing new competitive actions and its capability and motivation to act on the opportunities.

They tested the hypotheses on a sample of firms from the global automobile industry, about 12 major automakers over 16 years from 1988 to 2003. They find that network structure rich in structural holes has a positive direct effect on firms’ ability to introduce a greater number and a wider range of competitive actions. However, the effect of dense network structure is contingent on firms’ IT-enabled capability. Firms benefit from dense network structure only when they develop a strong IT-enabled capability. Our results suggest that IT-enabled capability plays both a substitutive role, when firms do not have advantageous access to brokerage opportunities, and a complementary role, when firms are embedded in dense network structure, in the relationship between network structure and competitive actions.

Week 2 Reading Summary – Tambe et al.(2012) – Xi Wu

Tambe, P., Hitt, L. M., & Brynjolfsson, E. (2012). The Extroverted Firm: How External Information Practices Affect Innovation and Productivity. Management Science, 58(5), 843–859.

Falling internal communication costs and new internal information practices enable a firm’s external focus, which is the ability of a firm to detect and therefore respond to changes in its external operating environment. The benefits from IT are valuable only with appropriate changes in decision rights and organizational practices. This paper argues that the combination of external focus, changes in decision rights, and IT investments form a three-way system of complements resulting in higher productivity levels.

The organizational practice measures are generated by a survey that was administered to 253 senior human resource managers in 2001. IT employment data are obtained from a large sample of U.S.-based IT workers. Capital, labor, research and development expense for the firms are using the Compustat database.

Since providing direct evidence of complementarities is challenging because of endogeneity problem, the existing literature on organizational complements has therefore focused instead on providing evidence of the economic implications of complementarities between organizational practices. In this paper, Tambe et al. first find that external focus is correlated with both organizational decentralization, and IT investment. Second, they find that a cluster of practices including external focus, decentralization and IT investment is associated with improved product innovation capabilities and high productivity.  Moreover, when these complements are included in a production model, main effect estimates of IT and other organizational factors essentially disappear, indicating that firms derive the most benefits from implementing the system of technological and organizational resources, not IT alone. Third, this paper introduces a new set of instrumental variables and find that the results are robust.

Week 2 Reading Summary – Tae et al. (2018) – Xi Wu

Tae et al. (2018) When Your Problem Becomes My Problem: The Impact of Airline IT Disruptions on On-Time Performance of Competing Airlines

Disruption within the competitive environment and its subsequent effect on firm performance has been a crucial topic, yet there is few research about the effect on the competitive environment on rival firms when a resource-sharing firm experience an operational disruption. As is well established, the disruption of organizational routines often leads to performance degradation.  This paper argues that, in the presence of a shared resource, the disruption of the focal firm’s routines will also lead to the disruption of the competitors, and subsequently diminish the competitor’s performance. In addition, following the idea of routines as coordination mechanism, the study also examines how the complexity of routines, of both the disrupted firm and the competitor, affects the ability to react.

The empirical context is U.S. airline industry, where individual airports are the shared resource. Airlines have been subject to multiple recent large-scale disruptions. This paper examines the change in on-time performance (OTP) of airlines with and without IT disruptions, for flights departing from and arriving at hub airports of the airline experiencing IT disruptions, relative to all other flights. Complexity of routines could be indicated by full-service carriers (FSC) and low-cost carriers(LLC).

Using a sample data of 75,051 flights during four IT outrage incidents, this paper conducts fixed effect empirical models, and find that disruptions to one airline may have negative effects on the on-time performance of competitors who share the same resource. The direction and the magnitude of the effect is conditional on the routine complexity of the firms. This study complements the literature on firm adaptability under environmental change by considering the role of firm interdependence, and shed light on the impact of routine complexity on firm adaptability and performance.

Week 1 Reading Summary – Ho et al. (2017) – Xi Wu

Ho, J., Tian, F., Wu, A., & Xu, S. X. (2017). Seeking value through deviation? Economic impacts of IT overinvestment and underinvestment. Information Systems Research, 28(4), 850-862.

Information technology (IT) investment decisions are of continuing strategic importance to firm. Under Red Queen competitive pressures, each firm’s performance depends on exceeding the actions of its rivals. This logic is useful but underresearched in the context of IT investment and its economic impacts.

This paper addresses the economic impacts of information technology overinvestment and underinvestment decisions. Hypothesis in this work draws on the view of Red Queen competition in conjunction with institutional theory. It argues that there is an effect of deviation on competitive advantage and an effect of deviation on legitimacy. Combining the two effects, a nonlinear relationship between firm performance and over- or under investment is expected (Hypothesis 1). In addition, drawing on the idea of management control mechanisms, the nonlinear firm performance impacts are conditional on ownership concentration (Hypothesis 2).

The authors use Tobin’s q to indicate firm performance, firm IT investment deviation is the difference between a firm’s IT capital and Industry IT. Sample set covers S&P 500 firms during 2001–2006. Firstly, ordinary least squares regression is applied. Then, to handle the possible endogeneity problem caused by selection bias, this paper uses Garen’s two-stage approach to model it. Another two partial adjustment models are applied to help address the dynamics of IT investment.

From the four models result, this paper finds that, on average, there is a positive relationship between a firm’s overinvestment in IT and Tobin’s q, although the relationship attenuates at higher levels of over-investment. However, on average, there is no relationship between a firm’s underinvestment in IT and its Tobin’s q. In addition, family ownership positively moderates the performance impact of underinvestment. These findings partially support hypotheses proposed.