Dynamic Portfolio Theory and Management (Hardcover)

Intuitively, we know that the market responds to changes in economic inputs. The trouble is, we don't know which inputs carry the most weight, which ones lead or lag and when they are the most powerful. Dynamic Portfolio Theory is the first book to look at many studies by other research professionals in order to sort through the clutter of hundreds of indicators and rank the indicators according to their prominence in these other studies.

Oberuc's approach to portfolio construction removes the problems of trying to determine expected rates of return, co-variances and volatilities. He has developed a model called Dynaporte that has the distinct advantage of being able to respond to changing market conditions as they occur by linking the asset allocations to changes in macro-economic variables. Although the book is not intended for math-phobic readers, it supplies sufficient theory to satisfy inquiring minds. However, for those of us without advanced statistics proficiency, the conceptual discussions provide plenty of meat to chew on.

He questions each of the Markowitz assumptions and compares their value in predicting portfolio performance to those used in the DynaPorte model, a multi-variate linear regression technique technique based on macro-economic and market based inputs. For example, the clear explanation surrounding the relative merits of mean variance, downside variance and mean absolute deviation as risk measures is highly enlightening. This and similar examples alone make the book worthwhile, as so many investors have used the MPT model without understanding why they are doing so.

Separate chapters are devoted to indentifying and developing the best indicators for stocks, bonds, interest rates and pieces of hedge funds. Starting with perhaps 20 inputs, the DynaPorte model reduces down to the best ones, perhaps 5 or 6 of the most statistically significant. Unlike neural networks, which take multiple inputs and screen them in mysterious ways to derive those which track historic performance most closely, the DynaPorte model shows why certain independent variables are being retained and others discarded. Each variable is scored by its t-score, R-squared and other measures.

What makes the process fit together is the integration of each of the individual models into time-sensitive efficient portfolios. Oberuc devotes a full chapter to the recent literature studies discussing both the difficulties of getting good inputs for data and the tests for predictability of market returns. Again, a chapter worth the price of the book.
Whether it's called market timing, tactical asset allocation or chopped liver is not really important. What counts is whether the historic in-sample results translate into out-of-sample portfolio benefits. In this case, it's enough to make a market timer weep tears of joy and wear the label with pride.

In-sample optimization of a stock, bond and t-bill portfolio from 1980 through 2001 showed a 5% annual excess return over a pure stock portfolio with substantially lower volatility. Another study, using Fidelity sector funds did even better. Both studies accounted for transaction costs. Out of sample results were comparable. More remarkably, the portfolios continued to rise during the post-bubble period.

It is important to note that the results discussed in the book did not just fall out of the massaging of the inputs. Many different combinations were tried over many months before coming up with a satisfactory model that fit the market's behavior. The techniques described are not a magic bullet. However, with increasing interest in tactically using index funds and multiple style portfolios, those tools able to handle a dozen or more market sectors rapidly deserve attention.

Whether the reader can construct portfolios like those achieved in the book is not why it should be bought. The real benefit is in the full-featured discussion of all of the factors that go into making a robust modeling system and as a significant contribution to modeling literature. Consultants and portfolio managers need to give this book a place on their shelf.

The first part, from Chapter 1 to 9, contains a concise survey of the academic innovations in portfolio management in the past five decades, with the author's useful insights on these models and evidence. In particular, it emphasizes on the idea of dynamic portfolio allocation, which uses broad economic information to actively manage investment allocation. This part contains not only elegant models but also insightful intuition, so that this book becomes readable to virtually anyone interested in portfolio investment and management. So I think it can be an excellent reference for both doctoral students and financial practitioners in this field.

The second part advocates DynaPorte investment model that links asset allocations directly to economic factors. DynaPorte represents innovations of incorporating economy-wide information into your portfolio formation, which is the essence of dynamic portfolio management. Moreover, the recent academic evidence against the Random Walk Hypothesis has opened the door to superior long-term investment returns through such kinds of active investment management. I believe that the DynaPorte can help improve the practitioners' understanding of dynamic investment and thus make better use of various kinds of information.