Detecting the Madoff Effect: Methodology for Fraud in Hedge Funds

As a result of the recent Bernard Madoff fraud scheme, pension funds and corporate finance managers have been put on the defensive, wondering how to detect this type of “under the radar” deceptive scam. The depth of the fraud in the case of Bernard Madoff and his ability to engage in a $50 billion undetected scheme employing “serial correlation” demonstrated the vulnerability of financial institutions to trusted individuals operating inside the security model. Bernard L. Madoff Investment Securities LLC (“Madoff”) engaged in a ponzi or pyramid fraudulent scheme in which investors were paid interests not from actual investments but from the funds deposited by other investors. But being on the defensive is not the ideal solution, as it forces financial institutions into a reactive mode, always trying to catch-up with the perpetrators, who somehow remain one step ahead. We would like to suggest a more proactive approach and corresponding methodology for detecting fraud in hedge funds.

Madoff was able to hide his scheme using a “serial correlation” reporting scheme. A serial correlation is a term used by MIT professor and hedge fund theorist Andrew Lo to describe the degree to which each month's returns in a fund mirror the results of the month before. Dr. Lo’s theory is that is a hedge fund has a nice smooth line in its rate of return every month. Upon close examination, any variation to the “smoothness” of the line constitutes a red flag, which should be look upon more carefully.

In the last year corporate finance departments, financial institutions, as well as public and private pension fund portfolios have already lost about 33% of their values due to overleveraged investment banks, the housing and credit crises. An effective, (no more than 1 to 3 weeks) and cost efficient proactive methodology to detect the Madoff effect in the hedge funds would be to apply the following methodology in the specified order to the relevant data available to you:

• Link Analysis – Use link analysis to determine in the network Madoff is categorized by, and create a subset of that network of hedge funds.
• Predictive Modeling – Use predictive modeling to create a score of all the hedge funds in your subset. Use Madoff’s variables as your training data.
• Clustering Analysis – Perform a cluster analysis which includes among other variables the predictive score. Since the predictive score is a multidimensional variable when used with one-dimensional or “flat” variables you will obtain a binocular vision (or binocular summation) of your analysis and increase by 1.4 times the ability to detect the serial correlation. See, Improving Search Engine Optimization by Incorporating Predictive Analytics at http://atomai.blogspot.com/2008/12/improving-search-engine-optimization-by.html
  
  For verification of the analysis you could use the following factors:
• The reputation of the independent auditors of the hedge fund identified through this methodology;
• Control Chart using standard deviation of the yearly returns over a 3-5 year period (exclude the current year);
• The ratio of total number of employees to the total amount of investments.
  
  Contact Alberto Roldan at atomanalytics@gmail.com or Sean Suskind at seansuskind@gmail.com
  

Improving Search Engine Optimization by Incorporating Predictive Analytics

As more companies increase the size of their databases search engine optimization (SEO) techniques can be adapted to data mining of commercial databases. In SEO link analysis is a measure of the quality and relevance of the set of links pointing to a given site. This is measured is achieved through an algorithm that maps the hyperlinks in a series of networks. The measurement creates a ranking of the strength of the inbound links to a particular network. The objective of link analysis is to detect patterns or trends that would make the search engine to bring to the top the most relevant web sites in any search.

Link analysis contains multiple variables that are analyzed. Google’s claims over 200 variables are analyzed in its link analysis for its ranking algorithm. Although I do not know which variables it uses, I surmise that they are the keystone of Google’s success. The core of any analysis is its variables. Let me suggest the utilization of predictive modeling as an additional variable that will improve SEO.

Using predictive modeling as another variable in link analysis could potentially increase SEO by 1.4 times by giving depth perception to the link analysis. In ophthalmology medicine it an established fact that binocular vision gives depth perception, and that depth perception (or binocular vision) increases the range of view 1.4 times greater than monocular vision. In other words, you can see better with two eyes than with one eye. The equivalent of depth perception in analytics is the addition of a predictive modeling (or scoring) variable to any pattern detection analysis.

A predictive modeling variable will improve the SEO because:

1. It gives an independent variable that acts as a spare variable in case that another variable is not working. In other words, you can use a predictive modeling variable in a correlation analysis as your independent variable against the other numerical variables in your link analysis.
2. A predictive modeling variable will widen the field of view of your networks from 160 degrees to 200 degrees.
3. Binocular summation (seeing with two eyes) will enhance faint but important networks and links within your data.

Among SEO scientists, statisticians, and business analysts it would increase stereopsis, or the keen sense that they have depth perception. In other words, it would give them another tool to do their work more efficiently.


Most of the variables used in link analysis are flat, or with one-dimension. A predictive modeling variable is multidimensional and hence a “depth variable”. The addition of a “depth variable” to any analysis statistically can be expressed as detecting the networks using two sensors instead of one. If each flat variable alone had a 0.6 probability of detecting a network, that probability has been calculated to be:

Pb = Pr + Pl - (Pr x Pl) = 0.6 + 0.6 - (0.6 x 0.6 ) = 0.84 (1)

The improvement from 0.6 to 0.84 represents a 1.4 fold improvement. This improvement can be achieved in any analytics technique by adding a multidimensional variable to a one- dimensional variable during analysis.

Contact: Alberto Roldan, CEO of R&R Analytics at atomanalytics@gmail.com

Depression Economics: America’s Economic Crisis

I have written many times in the last 12 months about this issue in the Business Analytics website at http://atomai.blogspot.com/. This interview from Paul Krugman in Newsweek gives us an insight into how deep is the ocean in the financial and credit crisis. If you look in page 2 of the interview Krugman talks about a $10 trillion (yes trillion with a “t”) shadow banking system that just went up on smoke. Hence, he argues that a $700 billion capital injection by the Federal Reserve and Treasury Department will be insufficient to make up for those $10 trillion. I agree with Krugman, but we need to consider that if we take into consideration the capital injection that have an effect in America’s economic crisis by other countries in Western Europe, Russia, China, and Japan. We are talking more than $3 trillion injected into the world economy when we take those capital injections into consideration. This still leaves us with an $7 trillion issue. The next issue is the valuation of those depleted assets, because it is the difference between the $7 trillion and the value of the depleted assets that are going to determine when we are going to touch bottom. I estimate that the minimum value of those assets will be about $3.6 trillion, and that is going to leave us with about $3.4 trillion that the international capital markets are going to deal with a combination of global stock market devaluations and injection of additional capital by central banks. The jobs program and other programs by president-elect Obama will have an effect on the economic crisis but it is too early to determine what that effect will be. If the administration of President-elect Obama injects into the economy $1 trillion through different programs, including a jobs program, we will still have $2.4 trillion to deal with. The bottom line is that companies need to prepare to cut 12% to 24% in additional expenses above what they have already trimmed. The role of cost-efficient business analytics decision-support systems at the operational level is going to become a cornerstone in the transformation of companies in how to increase the margin of profit with less resources.

The Krugman interview is at http://www.newsweek.com/id/171871/page/1

Successful Business Intelligence Projects: The Role of Managers and Leaders

Most BI projects fail because the leadership for those projects is wrong. An article on how Some Brains Wire for Change[i] helps explain the physiological reasons how some individuals can adapt easier to change than others. This article makes clear that people’s brains are different and that different does not mean “bad”. In today’s recession it is important for organizations to understand the role of a manager vis-à-vis the need for a leader in the area of analytics.

Every organization needs both managers and leaders in analytics. Managers are those individuals who supervise individuals who conduct analytics within an organization. Leaders are those individuals who guide or have commanding authority in the area of analytics within an organization.

I have found that organizations tend to have good managers in the area of analytics but lack leaders. Managers are efficient at maintaining the status quo and are adverse to risks. Leaders are risk takers and innovators, but not necessarily proficient at managing or maintenance of a department.

An organization that is satisfied with how its analytical capabilities are producing a lift in their revenues and profits, should be looking to improve how to efficiently manage those capabilities. On the other hand, an organization that is looking to improve revenues, costs, or profitability by using its analytical capabilities needs leadership in the area of analytics. A good manager realizes when he needs a leader, and a good leader acknowledges the need for a manager.

Managers and leaders of analytics have different roles, and although they are not mutually exclusive it is the role of executive management to define the priorities in the area of analytics at any given time. Sometimes organizations make the mistake of trying to make managers leaders or vice-versa. The results are that the analytics capabilities within an organization never bloom to its full potential in contributing to increased profits. In my experience managers contribute about 80% to 90% of the success of a business intelligence project, and leaders contribute 10% to 20% of the success of the project. Therefore, a successful BI project needs both managers and leaders.

Contact Alberto Roldan at R&R Analytics at atomanalytics@gmail.com
[i] http://www.livescience.com/health/081201-brain-personality.html