«VOLUM E 1 1, N UM B E R 1 I S SN 2 1 6 8 - 0 6 1 2 F L ASH DR I V E I S SN 1 9 4 1 - 9 5 8 9 ON L I N E T h e In s t it ut e f o r Bu s i n e s s an ...»
G L OBA L C ON F E R E NC E ON
B U S I N E S S A N D F I NA N C E
VOLUM E 1 1, N UM B E R 1
I S SN 2 1 6 8 - 0 6 1 2 F L ASH
DR I V E
I S SN 1 9 4 1 - 9 5 8 9 ON L I N E
T h e In s t it ut e f o r Bu s i n e s s an d F i n an c e R e s e arc h
Ho n o l u l u, Haw ai i
Janu ar y 4 - 7, 2 0 1 6
TABLE OF CONTENTS
ENGLISH PROCEEDINGS 1
SPLICED CORRELATION: THEORY DEVELOPMENT 2Jeffry Haber, Iona College 2
CHINESE ENTREPRENEURS IN SMALL AND MEDIUM ENTERPRISES (SMEs) – AN EMIC VIEW 5Yunke He, Okanagan College 5 Heather Banham, Okanagan College 5
PUBLIC FINANCE, MICRO FINANCE AND ACCELERATED ECONOMIC DEVELOPMENT FOR THE
ERADICATION OF EXTREME POVERTY IN SUB SAHARA AFRICA 10Chiaku Chukwuogor, Eastern Connecticut State University 10
FORMATION OF BRAND PERSONALITY ASSOCIATIONS: THE ROLE OF PRODUCT CATEGORY
INTEREST AND CONSUMERS’ PROACTIVE PERSONALITY 18Ove Oklevik, Sogn og Fjordane University College 18
INTENTIONS TO USE AN ONLINE RESTAURANT REVIEW WEBSITE 23Joshua Fogel, Brooklyn College 23 Mohit Kumar, Brooklyn College 23
CAREER STRATEGIES OF HOTEL MANAGERS IN CANADA 29Candace Blayney, Mount Saint Vincent University 29 Karen Blotnicky, Mount Saint Vincent University 29 U.S. CORPORATE PENSION EXPENSE AND THE 2007-2009 FINANCIAL CRISIS: AN INTERRUPTED
TIME SERIES ANALYSIS 38Benjamin B. Boozer, Jr., Jacksonville State University 38 Julie A. Staples, Jacksonville State University 38 S. Keith Lowe, Jacksonville State University 38 Robert J. Landry, III, Jacksonville State University 38
A MODEL FOR GENERATING AND SHARING INFORMATION ABOUT CULTURAL ASSETS 47Ece Zeybek, Istanbul Arel University 47 Uğur Yozgat, Marmara University İstanbul 47 Meltem Gürünlü, Istanbul Arel University 47 LEADERSHIP
ENGLISH PROCEEDINGSGCBF ♦ Vol. 10 ♦ No. 2 ♦ 2015 ♦ ISSN 1941-9589 ONLINE & ISSN 2168-0612 USB Flash Drive 1 Global Conference on Business and Finance Proceedings ♦ Volume 11 ♦ Number 1
Correlation involves two data streams. Often a significant correlation relationship (uncorrelated, positively correlated or negatively correlated) in the long-term is not present in the short-term. Worse, often the shortterm correlation is contradictory to the long-term. Utilizing three sets of data, where two are combined into one at varying points of time could allow the long-term correlation to be also replicated in the short-term.
There remain various obstacles to overcome, such as scaling, determination of inflection points and the selection of the data streams, but this paper puts forth the theoretical justification for the concept.
Consider this thought experiment – you are a secret agent and you wonder whether someone is following you. You constantly look back to see if there is a discernable pattern of the people who are behind you. You cannot detect one. Now think of it from the pursuer’s perspective – assuming your paranoia was wellfounded. Knowing that you would be able to detect a consistent presence behind you, they alternate who is following you. Every other time you look back you see a different person. A trained spy might be able to adjust and detect the tail, but probably not an ordinary citizen. Correlation is not that much different from
the thought experiment. Consider the following streams of figures (noted “A,” “B” and “C”):
A B C
The correlation between A and B is 0.0; perfectly uncorrelated. But break the ten item string into two strings of five items each, and a different story emerges – the correlation of the first five items is +1.0, perfectly correlated, and the correlation of the last five items is -1.0, perfectly negatively correlated. The two strings of perfect correlation combine to produce perfect non-correlation. Now take the correlation between A and C – it is 0.5. If you take the correlation of the first five items you get 1.0, and if you take the correlation of the second five items you get 1.0. Each of the shorter periods is perfectly correlated, taken as a whole the correlation is halfway between 0.0 and 1.0. Now take the correlation between B and C – it is 0.0, and the two strings (of five items each) replicate the five items strings of A and B (+1.0 for the first five and -1.0 for the second five). Essentially what is being demonstrated is that the long-term correlation (ten observations) is valid as long as you hold both streams for the full duration. If you enter or leave an investment anywhere within the ten periods your correlation experience might be quite different.
Traditionally, in calculating correlation we take items two at a time over some period of time. Correlation is useful when it works, and not useful when it doesn’t. A significant correlation figure over longer periods may not be present when taken in shorter intervals, and vice versa. Correlation is not causality, but an indication of what we expect to happen to something by virtue of observation of something else. One of the limitations of correlation has been the need to find two streams of data (in an investment context, generally monthly returns or monthly closing prices) that are correlated in both the short-term and the long-term.
GCBF ♦ Vol. 11 ♦ No. 1 ♦ 2016 ♦ ISSN 1941-9589 ONLINE & ISSN 2168-0612 USB Flash Drive 2 Global Conference on Business and Finance Proceedings ♦ Volume 11 ♦ Number 1 Long-term correlation is desirable because of the reliance that can be placed on it over a long duration. But this is only true if the correlation exists when taken in shorter periods as well. It would be unusual for two streams of data that are highly correlated when calculated over a 15 year period (for example) to have noncorrelation or negative correlation over meaningful, but shorter periods (see Haber 2012, Haber and Braunstein 2009, Haber and Braunstein 2008). Likewise for non-correlated streams of data over the longterm – it would not be unusual for these streams to be highly correlated in shorter, but significant periods.
So, what is necessary then, is for there to be significant correlation (positive, negative or non-correlation) over the long-term AND the short-term. This has typically been hard to find – compromises in correlation have been accepted, meaning that shorter periods of aberrant behavior have to be tolerated for what is considered the greater good – delivering the long-term correlation. This paper develops whether there are any alternative means to find both long- and short-term correlation with no compromise.
Finding correlated streams usually involves selecting a group of variables and testing them two at a time.
Spliced correlation starts out the same way, but considers whether various streams can be combined to produce a long-term correlation as well as a short-term correlation. Consider a data stream, denoted as “X”.
This stream will be held constant – the second stream will be spliced to produce the hybrid stream of data.
Visually, it resembles Figure 1.
Figure 1: Depiction of When Splicing Would be Valuable
Stream “Y” correlated with “X” for a period of time, then starts into a period of not correlating. When “Y” is no longer correlating with “X” stream “Z” starts to correlate with “X.” Then when “Z” no longer correlates with “X” “Y” starts resumes its correlation. This alternates on a predictable pattern. Splicing the two streams together (“Y” and “Z”) produces a combined stream of data that can be used in correlation calculations with “X,” as depicted in Figure 2.
Figure 2 :Depiction of What the Spliced Sequence Would Look Like
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXCorrelation calculations are based on streams of numbers, no matter whether they are untouched, adjusted or an artifact of creation. The math will work just as well. What is not trivial is identifying the streams that will be spliced, and the inflection points where the splicing takes place.
To test the theory a narrative needs to be developed about how two data streams relate to each other, and when the circumstances under which they won’t. Then an additional stream needs to be identified that will work when the first doesn’t, and then the development of a system for understanding when the substitutions should take place. Because the calculation of correlation is highly sensitive to sign changes, implementing GCBF ♦ Vol. 11 ♦ No. 1 ♦ 2016 ♦ ISSN 1941-9589 ONLINE & ISSN 2168-0612 USB Flash Drive 3 Global Conference on Business and Finance Proceedings ♦ Volume 11 ♦ Number 1 a substituted set of data might introduce sign changes, simply because the figures are based on a different scale (think of using stock quotes for stream “Y” and US Treasury bond yields for stream “Z” – because they are denominated differently, any change from Y to Z will introduce a sizable decrease in data (and likewise there will be a corresponding large increase in going to Z to Y). The relative scaling of the two streams is a consideration. How to determine the inflection point is another complication. Whether substitutions will be based on an ad-hoc model, where based on some algorithm the system will determine when to substitute or whether the substitution will take place on a scheduled basis (like every 5th day, for instance) will need to be worked out through testing.
Correlation is most valuable when it works in the short-term as well as the long-term. Historically, longterm, significant correlations go through extended short-term periods where the correlation is different from the long-term, in a non-trivial manner. Non-correlation can become extreme correlation; highly negative correlation can become highly positive correlation. Since correlation is based on two streams of data, there is an opportunity to combine two streams into one, for the purposes of calculating correlation. The combination of two streams into one should the long-term correlation to be replicated in any short-term period, given that the additional stream is correlated similarly during short periods when the first stream is not, and the model knows when the appropriate points occur when substituting back and forth.
REFERENCESHaber, Jeffry “Fooled by Correlation: How Blind Acceptance of Correlation Dogma Destroys Diversification,” Journal of Business Diversity, Volume 12(3), 2012, pp 22-25 Haber, Jeffry and Andrew Braunstein, “Examining the Role of Short-Term Correlation in Portfolio Diversification,” Graziado Business Report, 2009, Volume 12, Issue 3 Haber, Jeffry and Andrew Braunstein, “Correlation of Uncorrelated Asset Classes,” The Journal of International Business and Economy, Volume 9, Issue 2, December 2008, pp 1-12 GCBF ♦ Vol. 11 ♦ No. 1 ♦ 2016 ♦ ISSN 1941-9589 ONLINE & ISSN 2168-0612 USB Flash Drive 4 Global Conference on Business and Finance Proceedings ♦ Volume 11 ♦ Number 1
In-depth personal interviews are conducted with nine Chinese entrepreneurs. They are located in three different geographical regions from coastal to interior. Their businesses span eight sectors from clothing to industrial automation. This article focuses on their motivations to start new businesses, and introduces their educational and working background as well. The interview results show that eight out of nine entrepreneurs resigned from their stable or even enviable positions to pursue their personal dreams - from creating own brand in men’s suits to commercializing research findings from own PhD dissertation in industrial automation. The research findings have some important implications on engineering education, female entrepreneurship, and forms of business ownership for SMEs.
JEL: M190 KEYWORDS: Chinese Small and Medium Enterprises (Smes), Chinese Entrepreneur
INTRODUCTIONThe increasing importance of Small and Medium Enterprises (SMEs) is a global phenomenon brought about by market forces, technological advances, personal career aspirations and underlying demographic changes in the population. In the Organization for Economic Cooperation and Development countries, SMEs account for 95 percent of the enterprises and 60 to 70 percent of the employment. The transformation of the economy in China is also relying on the role of entrepreneurship and the development of a strong SMEs sector. SMEs are drivers of employment, innovation, exports and gross domestic product. Despite the importance of SMEs to the economies worldwide and their resilience to economic downturns, there has been a shortage of data concerning SMEs which is partly attributable to the fact that it is a very challenging area for researchers (Curran & Blackburn, 2001). Country differences including those in relation to the role of government and other independent bodies such as banks impact the establishment and growth of SMEs (Uhlaner, Wright & Huse, 2007). It is also extremely challenging for researchers to undertake research in China due to language and cultural barriers. This research project intends to contribute to the body of knowledge on SMEs with this exploratory research on SMEs in China.
SMEs provide investment opportunites and attract the attention of policy makers to foster growth and increase resilience in national economies. This is also the case in China where SMEs are significant contributors to economic growth, employment creation, exports and techonology innovations (Chen, 2006).
While an increasing amount of research is being undertaken on SMEs in China (Cunningham & Rowley,