Artificial Intelligence is a mathematical or logical procedure used to solve complex decision problems. In Artificial Intelligence, the procedure learns solutions to problems from large quantities of historical data. An example of artificial intelligence is Facial Recognition: the use of stored facial measurements of known individuals to later identify them from facial imagery. For Chapoquoit, we employ monthly data back to the 1970’s on the returns of multiple investments and the values of a set of macroeconomic factors to deduce the optimal way to dynamically change asset allocations.

Linear Programming is a mathematical procedure that optimizes a linear objective function while satisfying a large number of linear constraints. An example of the use of linear programming is the process by which an oil refinery is controlled to produce needed minimal percentages of gasoline, jet fuel, heating oil and other distillates in order to maximize the profit from the refinery. For Chapoquoit, the changing allocations must sum to 100% for every month of history in a way that produces an historical target return and minimizes the sum of portfolio losses.

Macroeconomic Factors used by Chapoquoit are measures of the economic or market environment that are logically related to the performance of investments. Examples of macroeconomic factors include the unemployment rate, 10-year government bond yield, S&P500 dividend yield, manufacturing capacity utilization, and the interest rate credit spread. Chapoquoit’s set of 20 macroeconomic factors was determined through extensive research. This proprietary set of factors directly impacts our performance and marketing edge.

Chapoquoit Employs 30 ETF Investments that represent a range of U.S. equity sectors, foreign equity sectors, fixed income sectors, and several commodities. Using this broad set of investments can be classified as a global macro approach.

Dynamic Asset Allocation Control Functions are linear functions of macroeconomic factors used to control the time-changing asset allocations. Each investment in the Chapoquoit model has its own asset allocation control function. For example, the asset allocation control function for a 10-Year Government Bond ETF might look like:

10-Year Government Bond ETF Allocation = α + β1 (Unemployment) + β2 (10-Year Bond Yield) + β3 (S&P500 Dividend Yield) + … βn (Factor n)

Where the unknown coefficients (α, β1, β2, β3 … βn) are determined simultaneously for every investment through the use of linear programming. This is done in a way that produces an historical target return and minimizes the sum of portfolio losses.

The Portfolio Risk Measure employed by Chapoquoit is the average value of the portfolio loss in those months when the portfolio suffers a negative return. This is a true downside-risk measure. The optimization process seeks to keep this measure to a minimum.