The Journal of American Business Review, Cambridge
Vol. 6* Number 1 * December 2017
The Library of Congress, Washington, DC * ISSN 2167-0803
Online Computer Library Center, OH * OCLC: 940146916
National Library of Australia * NLA: 9026139
Peer-Reviewed Scholarly Journal
All submissions are subject to a double blind peer review process.
The primary goal of the journal will be to provide opportunities for business related academicians and professionals from various business related fields in a global realm to publish their paper in one source. The Journal of American Business Review, Cambridge will bring together academicians and professionals from all areas related business fields and related fields to interact with members inside and outside their own particular disciplines. The journal will provide opportunities for publishing researcher's paper as well as providing opportunities to view other's work. All submissions are subject to a double blind peer review process. The Journal of American Business Review, Cambridge is a refereed academic journal which publishes the scientific research findings in its field with the ISSN 2167-0803 issued by the Library of Congress, Washington, DC. The journal will meet the quality and integrity requirements of applicable accreditation agencies (AACSB, regional) and journal evaluation organizations to insure our publications provide our authors publication venues that are recognized by their institutions for academic advancement and academically qualified statue. No Manuscript Will Be Accepted Without the Required Format. All Manuscripts Should Be Professionally Proofread Before the Submission. You can use www.editavenue.com for professional proofreading / editing etc...
The Journal of American Business Review, Cambridge is published two times a year, Summer and December. The e-mail: email@example.com; Website, JABRC Requests for subscriptions, back issues, and changes of address, as well as advertising can be made via the e-mail address above. Manuscripts and other materials of an editorial nature should be directed to the Journal's e-mail address above. Address advertising inquiries to Advertising Manager.
Copyright 2000-2017. All Rights Reserved
The Omniscience Model: Lagged Correlation
Dr. Jeffry Haber, Iona College, NY
Patrick Hardiman, Iona College, NY
The Omniscience Model is an attempt to harness the nearly limitless flow of real-time financial information with the geometric gains in computing power to predict future stock prices. A previous paper (1) reviewed various mathematical and statistical functions from a variety of fields to determine which would be applicable to stock price prediction. Correlation was one metric selected that had potential. Correlation is the relationship between two streams of data and can be any value between (and including) -1.00 and +1.00. The extremes (-1.00 and +1.00) represent perfect correlation, whereas values close to 0.00 indicate non-correlation. The sign (positive or negative) indicates whether the relationship is direct or inverse (respectively). In a direct relationship as one stream increases the other stream would increase as well. In an inverse relationship as one stream increases the other decreases. In an investing context correlation is important because it represents the inference that can be made by utilizing one stream of data to predict the value (or change in value) of a second stream. The problem with correlation is that it is not investable. Both streams of data are for the same temporal period, which means that by the time you have finished the calculation the ability to invest has passed based on the information provided by the correlation. This paper seeks to develop the concept of lagged correlation – that is, where one stream of data is for period t and is used to calculate the correlation with a second stream of data at period t+x for possible use as a filter in The Omniscience Model. Correlation is a ubiquitous metric in investing, usually used in the context of how a fund or security behaves in relation to an index, benchmark or sector. Many fund sponsors will tout that their fund is “uncorrelated” with stocks, which would position their fund as a risk diversifier for a portfolio. Since correlation is a widely-used statistic, it is not surprising that it has also been widely researched. Most specifications of correlation are for extended periods, such as 10 or 15 years. Studies have shown (2) that correlation calculated over the long-term might not hold in shorter segments within this longer term. The typical calculation is based on returns (as opposed to prices). There is sound statistical reasoning for this, but at the same time it is also potentially hazardous when the data looks like a horizontal (or vertical) line when plotted. (3) As conventionally used, correlation is a descriptive metric, which helps an investor understand how the fund or security reacts, given what has already happened to another fund, security, benchmark or index. To be usable as an investing metric correlation needs to be predictive rather than descriptive. A version of correlation is perfectly suited to this task. Commonly labeled “cross-correlation,” or as we prefer to call it “lagged correlation,” it involves off-setting on stream of data. Peterson states “Linear correlation coefficient as a function of time lag is the “cross-correlation function.” (4) It is easiest to describe correlation with a cause-and-effect narrative description. When it gets cold, people dress warmer. There is a direction of cause-and-effect, first it gets colder then people dress warmer. It doesn’t get cold because people dress warmer. If one data stream was temperature and the other was level of warm clothing warm, we would expect to see a high correlation. Some cause-and-effects do not cause the change that can be observed so close in time. Since weather forecasts are available for the next day, people could anticipate the level of clothing to wear so that cold temperatures are met with warmer clothing with nearly no lag. Rainfall and river flow are two data streams that would seem to have a cause and effect relationship with some lag. (5) Other cause-and-effect relationships that would seem to have a lag include television ads and the sale of a product and tweets about a product and sales (6) and pressure and state indicators and North Sea fish population. (7) While these examples are where a cause-and-effect relationship is a reasonable hypothesis, cause-and-effect is not a requirement to apply correlation or lagged correlation. For instance, if two companies compete in the same industry and are roughly the same size with similar operating characteristics, it would not a stretch to think that are population favors or disfavors their products their stock prices might move in similar fashion (think Pepsi and Coca Cola). The increase in Pepsi’s stock price would not cause the increase in Coca Cola’s stock price, but it is not crazy to think that both stock prices would move in the same direction if they are susceptible to similar market changes. For any two streams of data, one set becomes the predictor and the other the target. The target is offset by some amount of time, call it “x,” so that the predictor at time t will be used as one stream and the target stream will be based on t+x. The lagged correlation model (with a lag of k) is described by (8) When making a standard correlation calculation it does not matter which stream is first and which is second. When calculating lagged correlation, the order does matter. This paper does not develop which stream should be the predictor and which is the target, but performs two calculations for each pair, with the predictor and target switching in the second calculation. This paper is exploratory in nature and develops the process, validity and use, but not the efficacy of lagged correlation. The two company stocks chosen to represent the streams of data were AMC Theatres (ticker symbol AMC) and Sony Pictures (ticker symbol SONY). These two companies were chosen judgmentally – they are both one of the largest in their sectors, and it is believed that there is a relationship between a movie studio and a movie theater (one provides content and one provides venue). While this belief might be incorrect, it nonetheless provided the basis for the choice of the two companies. The companies will illustrate the process and determination of value will be up to the reader. For each company, one year of daily stock prices was chosen (the unadjusted close). The period of time was May 23, 2016 through May 22, 2017, which provided 252 observations. We are not suggesting the use of daily stock prices, but use them for convenience. From this point on we will refer to “periods” rather than “days.” We then took the correlation between the two streams, for all observations then in rolling periods. We did this for both prices (Figure 3) and returns (Figure 4). And we also did it using SONY as the predictor and AMC as the target and then with AMC as the predictor and SONY as the target. We used rolling 5 day periods, not based on any science but because it would yield more correlations. Further study should be done to develop an optimal rolling period. We then took lags of 0, 1, 2, 3 and 4 periods. A critical factor in any correlation analysis is “what is considered high (or strong) correlation” (for this paper) or “what is considered uncorrelated (or non-correlated)” in other contexts? There is no accepted framework for what is considered strong or weak or uncorrelated. A taxonomy that was developed, that we think is useful is (9) Since the square of the correlation is the coefficient of determination, which represents the explanatory power of the relationship, using the low end of the range for “extreme correlation” would be a coefficient of determination of .64 (.80 squared). We decided to use a cutoff of a correlation of .90 (which explains 81% of the variation). This feels (in an admittedly arbitrary manner) like a better cutoff than an r of .80 (where the explanatory power of 64% seems too close to 50%, which is even-odds in a binary decision).
American Depository Receipts: An Analysis of the Underlying Stock Returns
Dr. Chih-Chieh (Jason) Chiu, Rider University, Lawrenceville, NJ
This paper examines option listing effects on American Depository Receipts and their local counterparts. I analyze a sample of 42 option listings from 15 different countries and obtain negative price effects on the day after option listing. One possible explanation for the findings is that option listing relaxes the short sale constraints for the underlying securities. Do option listings have an impact on the underlying securities? According to Arrow-Debreu uncertainty theory, option listings move an incomplete stock market towards a state of equilibrium by increasing the number of assets to hedge against different states of nature. Black and Scholes (1973), however, see options as redundant assets and price them as such in their renowned options model. Consequently, numerous options models assume independence between stock and options market. Without a concrete theoretical direction, I examine the option listing effects on the underlying American Depository Receipts (1) (ADRs) and their local counterparts in the home country. I find a negative price effect on the first day post option listings. The negative abnormal returns are -0.66% and -0.64% for ADRs and their local counterparts respectively. Empirical research has shown price effects of option introductions on the underlying assets. The milestone paper by Conrad (1989) suggests that option introductions have a permanent effect on stock prices. Using a sample of 96 option listings for the period of 1974 to 1980, she finds option introductions have a positive price effect starting three days before introductions. She attributes the increases to the heightened demand for stocks from traders. The positive price effect from option listing, however, has been dampened by the introduction of options on the S&P 500 futures. Using a sample of 300 option listings from 1973 to 1986, Detemple and Jorion (1990) observe positive price effects and reduced volatility due to option listings. More importantly, they find a decline of positive price effect after the introduction of option on S&P 500 futures in April 1982. Their findings are consistent with market completion hypothesis that options market stabilizes financial markets. Researchers around the world seem to offer support for the findings in U.S. markets. Positive price effects around option listings are documented by Watt, Yadav and Draper (1992) for the U.K., Aitken, Frino and Jarnecic (1996) for Australia and Draper, Mak and Tang (2001) for Hong Kong. In contrast to the previous findings, Sorescu (2000) reports negative price effect for options listings that took place after 1981 that cannot be explained by either market completion hypothesis or regulatory changes. The author expands previous research on option listing effects by covering 2051 listing events from 1973 through 1995. He proposes that the negative price effect is due to option listing lowers the short sale constraint by reducing the cost of short selling that was traditionally done with the underlying securities. Mayhew and Mihov (2004b) test the robustness of Sorescu (2000) results and find negative returns after 1981. The negative trend, however, disappears when the authors use Conrad (1989) event study methodology to calculate the abnormal returns. The authors further investigate the short selling constraint hypothesis by examining trading volumes of stocks and options before and after the listings. They find no significant changes in volume before and after option listings. The lack of findings in Mayhew and Mihov (2004b) implies there is no increase in demand for the underlying securities from the option traders. American Depository Receipts (ADRs) have been excluded from previous studies on option listings because their underlying stocks trade outside the U.S. The fact that ADRs have underlying stocks, and perhaps options, trading in the home market does not necessarily mean that one could hedge the ADRs efficiently. For example, when covering a short position in an ADR with home market options, a trader would be exposed to additional currency risks that could increase the short selling costs. ADR options can be seen as a mechanism that allows for relaxation of the short selling constraints by lowering the short selling costs. My results show that the abnormal returns on the day after option listing are negative for ADRs and their local counterparts, which is consistent with the short sale constraints relaxation hypothesis. The data sample consists of option listings on ADRs from the period of 1981 through 2003. Out of the original 97 option listings, 23 listings have option listing dates fewer than 100 days apart from ADR listing dates and are excluded from the sample as their estimation windows are too narrow for any meaningful analysis. Additional listings were rejected due to data gaps or data unavailability that leaves us with a total of 42 listings in the data sample. The 42 listings represent 15 countries, including Chile, Finland, France, Germany, Ireland, Israel, Italy, Japan, Korean, Netherlands, South Africa, Spain, Sweden, Switzerland and United Kingdom. Companies from UK dominate the sample with 12 listings, followed by Japan with 6, and Netherlands and France each with 4. A total of six option listings took place in 1980s. The earliest listing in the sample took place in 1981 with Sony Corporation, and subsequent listings took place with Hitachi in 1982 and Honda in 1985. Three additional listings, Imperial Chemical Industry, Glaxosmithkline and Shell Transport and Trading Company, all took place in 1987. Since all option listings in the sample took place after the implementation of SEC-imposed moratorium(2) on new option listings in 1980 and 1981, the regulatory change has no ‘regime shift’ effect, as described by Sorescu (2000), on the ADR option listings. Moreover, the introduction of options on S&P 500 futures in April 21, 1982 also has no significant effect on ADR option listings because only two ADR option listings in the sample took place before the introduction of stock index options. All returns were calculated using Datastream Return Index that shows a theoretical growth in value of a share holding over time, assuming dividends are reinvested to purchase additional equity. Foreign exchange rates and local index data are also from Datastream. Preliminary analysis with Eventus and volume analysis were completed with data from CRSP. Using the event study methodology of Conrad (1989) and Sorescu (2000), I examine ADR returns after option listing. (See MacKinley (1997)) Two statistical models are used in computing the normal returns: the market model and the constant mean return model. Benchmarks are S & P 500 for the ADRs only analysis and major international indexes for the local stocks only analysis. No clustering effect is assumed and no portfolios were constructed because of the sparseness of the data over the sample period.
The Dead-Weight of Public Sector Companies in India: A Case Study
Dr. Anuja Gupta, Rutgers University, Camden, NJ
The Indian economy followed a socialistic model of economic development from 1947 to 1991, in which the government played a very large role in economic activity in the country. While liberalization started 25 years ago, there is still considerable dead-weight in the economy in the form of persistently loss-making public sector enterprises. Privatization, which was one of the planks of liberalization and a potential solution to this dead weight problem, has not made much progress in India. In this paper, we analyze the case of one such loss making public sector unit (PSU) - Mahanagar Telephone Nigam Limited (MTNL) with a view to gaining insight into the dire financial situation of many such PSUs. Further, we conduct a comparative analysis of MTNL with a private player in the same industry thus highlighting the glaring inefficiencies and poor performance of the PSU. We argue that there is no economic or strategic logic of carrying this dead weight in the economy. The assets tied up in these extremely inefficient companies could be put to much better use in India, a country with growth needs and ambitions. The Indian economy was following a “socialistic” model of government and economic development after its independence in 1947. This led to an economy that was replete with government control, and with direct participation of the government in economic activity through the ever-expanding public sector. In 1991, India faced a balance of payments crisis which led to a bailout by the International Monetary Fund (IMF) (Makhija, 2006). This, of course, came with stipulations to open up the economy, and get rid of excessive regulations. Thus, the process of liberalization started in 1991 with the government working on deregulation, delicensing and privatization. While reasonable progress was made in deregulation and delicensing, very little headway has been made in the area of privatization. As of 2016, there are still 244 public sector units (PSUs) under central government control. Only 14 PSUs have been fully privatized. In this paper, we highlight the dire financial position that many PSUs are in, by examining one such loss making PSU in-depth. We study the public sector telecommunications services provider Mahanagar Telephone Nigam Limited (MTNL). We argue that there is no strategic or economic imperative to continue operating enterprises like MTNL. Privatization is the only way to release the assets under control of these PSUs which can then be put to other productive uses in the economy. In the sections below, we first provide an overview of the current state of the public sector in India. We then present a case study on MTNL and discuss why privatization is the only way forward. In India’s socialistic or planned approach to economic development, five year plans were rolled out to define the objectives that the government would tackle in the five year window. The government was envisaged as playing a large role in setting up public sector enterprises in areas where the private sector would not (due to long gestation periods), could not (owing to the heavy capital requirements), or should not (owing to their strategic importance) participate. Examples of these were areas are infrastructure development, iron and steel, and national defense related industries, respectively. Thus, while there were 5 PSUs in 1951, this number gradually expanded to 244 in 2016 (see Figure 1 for overall growth of PSUs over the years. All Figures are in Appendix 1). Today, PSUs employ 1.4 million people, which is 4.8 % of the total workforce employed in the organized sector in India; they control 25 % of the total productive assets of the country, and contribute 14% to the country’s GDP1. This is a much larger presence of the public sector than many other market economies such as USA where employment in public sector was a mere 2% of total employment. Many of India’s PSUs operate in areas such as technology, tourism, and financial services that are not of strategic or national importance, and are left open to private players in most open economies. Loss-Making PSUs: In 2016, there were 244 operating PSUs. One third of these (78) were loss-making units. The total losses stood at Rs.28,756 crores ($4.4 billion), which translates into a loss of Rs.118 crores ($18 million) per operating PSU in the past fiscal year. The top ten loss making units account for 80% of these losses. Some of them are Steel Authority of India, BSNL, Air India, MTNL, Hindustan Cables, Hindustan Photo Films Manufacturing Ltd. Many of these are operating in industries where private sector players now dominate the market. Sick PSUs: Further, there are PSUs that have been declared “sick”. A unit is defined as sick if its losses in any financial year are 50% (or larger) than its Net Worth (averaged over the preceding four years). As of 2016, there are 65 sick PSUs that have been referred to the Board for Industrial Financial Reconstruction (BIFR). Although the reasons for their sickness vary, most of them face tough competition from the private sector and are burdened with obsolete plants and machinery, heavy interest on debt, surplus manpower and lack of adequate working capital. Of these 65 cases, only four have been approved for winding up operations. 48 PSUs were approved for revival by the Government or holding companies (till March 2014) envisaging total fund and non-fund based assistance of Rs. 40,937 crores. This comprises of cash assistance of Rs. 10,940 crores and non-cash assistance of Rs. 29,997 crores2. Some of the PSUs recommended for revival are Cement Corporation of India, Hindustan antibiotics Ltd, Bengal Chemicals and Pharmaceuticals. Given that all these activities can be performed successfully by the private sector (and indeed, are being performed successfully by the private sector in India) there is no clear logic or imperative to keep these sick units alive, by attempting to revive them. In fact, any attempts at revival is a drain of scarce public/budgetary resources. (like throwing good money after bad). PSUs with Negative Net Worth: Many PSUs have a net worth that is now negative– meaning that their accumulated losses over the years have eaten into the shareholder’s equity and retained earnings. These PSUs have effectively destroyed the equity capital of their investors. This is a very troubling situation of value destruction by for-profit businesses. There were 52 PSUs with negative net worth at the end of the year 2014. These represent the worst of the PSUs, which effectively have destroyed value over the years. PSUs versus Private Firms: PSUs compare poorly with the private sector in terms of performance. Between 1990 and 1998, the profit margin of public sector manufacturing firms was -4.4%, while it was 6.7% for private sector manufacturing firms (Department of Disinvestment, 2001). On the other hand, the cost structures of public sector firms is much higher than that of private sector firms. For example, wages as percent of sales in the same time window was 18% for public, and 8% for private firms. The wage bill of PSU workers are twice as high as in the private sector (Panagariya, 2008). Power and fuel as a percent of sales was 13% and 6.5% respectively. Let us now look at the case of MTNL to dive deeper and gain insights into some of these issues of financial malaise afflicting the public sector in India. Mahanagar Telephone Nigam Ltd. (MTNL) was incorporated in 1986 when the government of India merged the Delhi and Mumbai telephone networks. The area of operations of MTNL was restricted to these two cities, and the objective was to provide telecommunication services, expand the network and introduce new services. MTNL had a monopoly in telecom services in these two cities, with consumers in both cities having no choice in terms of service providers until 1991. Starting in 1991, and then in 2000 major policy changes took place in the telecommunications sector as part of India’s liberalization program. Private players, and then foreign players were allowed to enter the telecommunications market in India. MTNL operates under the administrative control of the ministry of Communication and Information Technology and the department of Telecommunications, with a 56.25% stake being held by the government of India. Business: MTNL provides fixed and mobile telephone services, GSM and broad band internet services in Delhi and Mumbai. MTNL launched its cellular services in Delhi and Bombay in 2001, and its broadband service in the two cities in 2004. It has two subsidiaries (wholly owned) - Millennium Telecom Ltd. (MTL) and Mahanagar Telephone Mauritius Ltd. It also has two joint ventures - United Telecom Ltd. (26.68% share) to provide CDMA based services in Nepal; and MTNL STPI IT Services Ltd. (50% share)3.