This is part three in a series describing the investment philosophy underlying our global, multi-strategy approach. If you missed the first two installations, presto:

• Part I: Elevator Pitch
• Part II: Philosophy
• You can also subscribe to our RSS feed if you like, in order to track new entries as we post them.

Now then…

We have mentioned that the Stracia process starts with certain assumptions about the means by which capital markets allocate resources. The purpose of this installment is not to explain our macroeconomic philosophy¹, but to describe how quantiative modeling fits into the overall investment process.

Quite simply, we start by testing our own assumptions. We construct and back-test the performance of thousands of hypothetical investment portfolios, using statistical and other “portfolio durability” techniques to determine the profit-potential of specific markets given our starting assumptions. We test the models’ results on a historical basis for general adequacy and ongoing statistical significance; and finally, use the results of these analyses to support decision-making in the here-and-now. So we do nothing without a tested, quantiative rationale for it.

One of the virtuous side-effects of our approach is that it breaks the investment process into discrete, manageable chunks: intermarket and macro-economic analyses support the asset allocation process; risk-based analyses of dozens of economic and market indicators support a long-term strategy for market entry and exit; and equity valuation models support individual stock-picking in both growth and value contexts — all for the purpose of improving risk-adjusted performance.

Calculemus

Each of the models submits thousands of data points to proprietary algorithms, which consist principally of proprietary computer code, optimizers and scenario-analysis procedures, and customized Monte Carlo and neural-network simulators.

Their nuts-and-bolts explanation, and our philosophy for software modeling in general, are beyond the scope of this article and inconsistent with the site’s objectives — which are to rationalize an intermarket portfolio strategy and to describe the strategy that we currently expect to yield superior risk-adjusted results. That said, we do occasionally elaborate on some fundamental considerations that go toward developing our software or its use, and may even post use-it-at-home code examples for users (in Excel spreadsheets or VBA, for example).

As noted, the data are typically time series: for example, historical market performance relative to some real-world trend, as measured by an economic indicator. The algorithms, in turn, apply many millions of calculations to the various data sets — including regression analyses and portfolio optimization routines — before finally submitting the results to a framework for coördinated long/short investment allocation among the stock, bond, commodity, currency, and derivative-products markets. The models support portfolio rebalancing with whatever periodicity is required, from daily adjustments to weekly, monthly, quarterly, etc.

A Thing Is What It Does: What Does Stracia Do?

The ultimate objective of the modeling process is to make readily actionable those data points that are truly meaningful, supporting a supple, responsive, and forward-looking decision process. Most single data points or releases will not have a major impact on portfolio construction, though many do inform allocation decisions at the margin.

As stated previously, the Stracia process encompasses the what of investing: a set of analytical, subjective techniques that are tested or qualified using objective modeling methods, supporting specific asset-category decisions at any point in time. However quantitative the approach, of course, it is still largely subjective, as we remain free to interpret the results of the models in the context of economic and market dynamics that, with each new cycle, are unique.

Therefore, Stracia also encompasses the why of investing: the underlying rationale (i) supporting a given asset-allocation stance and more generally, (ii) for expecting past relationships to influence future asset performance — or not. The models support Stracia’s explanatory or intellectual aspect. They justify the philosophy and make it actionable.

It is worth repeating that though it may be described as a “system,” Stracia is a system of thought — and its software engine, a framework for action — but not an expert system. Neither is it some sort of stock-picking doodad: Two independent users of the GOCMX, for example, may be expected to draw similar conclusions about appropriate asset allocation among investment categories, countries, and sectors. But they would also be expected to select different security baskets populating those strategies.² (We discuss investments and strategy but do not make actual investment recommendations here; see our disclaimer.)

In summary, this Web site describes the Stracia investment philosophy in the context of our models’ specific results. While Stracia is capable of expressing an intermarket portfolio-management strategy from the ground up, the reader can also use it to tweak her own strategy. For example, a manager considering increasing her exposure to foreign equity markets may employ Stracia’s techniques for eliminating certain kinds of political and country risk, the better to focus on those regions that Stracia determines most investment-worthy based on its well-defined, qualitative and quantitative geopolitical risk and elimination criteria (filters). We will have more to say on country risk in future posts.