Engineering Investment Research

The markets are too complex to simulate — We must rely on less perfect methods — Engineering research provides us with useful workarounds.

VIEW AS PDF

ENGINEERING INVESTMENT RESEARCH

We know we do not have the tools to accurately model public equity markets. The assumptions behind our current models are false, adjusting our models to reflect reality is too mathematically burdensome, and simulating the markets may be computationally impossible due to reflexivity.

Fortunately, scientists have faced similar challenges. Creative workarounds are typically required to arrive at breakthroughs on the cutting edge of science. After all, if our models and computing power were sufficient to solve a problem, the problem would be solved.

Richard Hamming, a mathematician, Manhattan Project contributor,  and member of Bell Telephone Labs, details a number of research methods in his book “The Art of Doing Science and Engineering: Learning to Learn.”

These workarounds have been used by successful investors for generations but are not often discussed.

The Method of Scenarios

As Hamming noted, "We have devised the method of scenarios to cope with many difficult situations. In this method, we do not attempt to predict what will actually happen; we merely give a number of possible projections."

We often hear that investors do not like to give predictions. But isn’t all investing a guess at the future? Hamming’s method provides a different view, differentiating between pure prediction and investing based on possible futures.

Successful value investors rely on Buffett’s vague advice: "Don’t lose money." Under the method of scenarios, Buffett’s admonition can be redefined. In our "number of possible projections," we must always include an extreme downside case. Our investment decision-making must prioritize this specific projection and, if in the downside case we are wiped out, we must avoid the investment regardless of potential return.

Garbage In & Accuracy Out

Accounting data is often garbage. Analyzing financials of both specific businesses and index aggregates creates the illusion of knowledge when none may exist.

First, the accounting numbers may simply be wrong, due to either fraud or honest mistakes. Next, accounting requirements, since they are designed for all businesses, often fail for specific businesses. (See Berkshire Hathaway’s reported results, which include the mark-to-market gains and losses on its investment portfolio as earnings, for an example of useless reporting.) Unadjusted financials are generally acknowledged as unrepresentative of reality, but adjusted numbers tend to be just as bad, with adjustments unfairly portraying results in a favorable light. Aggregating these issues across hundreds or thousands of businesses to generate benchmark data only compounds the problem.

Scientists to the rescue again! "If there is feedback in the problem for the numbers used, then they need not necessarily be accurately known." Investors can incorporate feedback, loosely defined here as conducting multiple analyses, to verify results.

Earnings power must be triangulated across all three accounting statements and verified via competitive benchmarking. When assessing intrinsic value, numerous valuation methods can be used to incorporate feedback, ultimately creating a mosaic that gives a blurry but directionally correct assessment of value.

Stability & Instability

Finally, without any modeling or simulation, we can construct portfolios to be robust to surprises. Stable systems (or anti-fragile, in Nassim Taleb’s parlance) can withstand shocks without breaking. We know from experience that financial markets are prone to shocks, so it is wise to consider this when constructing portfolios.

Many successful investment heuristics address this issue. Avoidance of leverage, higher-than-"optimal" cash holdings, and prudent diversification all contribute to portfolio stability.

The difficulty in implementation lies in the details. Wall Street frequently sneaks leverage into financial products unbeknownst to customers. Fund managers can easily take on above average risk when relying on deceiving metrics like net debt to EBITDA or interest coverage ratios. Diversification seems simple until correlations go to one.

Despite the challenges, though, portfolios can generally be evaluated as more or less robust without specific predictions of the future.

Better Tooling

It is generous to call economics a "dismal science." Borrowing methods from real scientists both highlights the failings of modern economics and provides better tools with which to approach the markets.

Not surprisingly, many of the heuristics used by successful investors reflect these actual scientific methods. While these methods lack the seductive allure of promised future returns, they are practical, robust, and implementable.