Changing your opinion when the facts change (or when Buffett buys Gold)

The Half-Life of Facts by Samuel Arbesman

Changing your opinion when the facts change sounds easy. But it turns out to be incredibly hard in practice.

I finished the Half-Life of Facts recently. It’s a quick read and discusses how facts change all the time. Samuel Arbesman is a mathematician and scientist. Nonetheless, his observations apply to almost every circle of life. Arbesman highlights multiple examples where knowledge and facts in every field evolves systematically. We can therefore generate reliable forecasts for how any field of study evolves. He observes that certain fields have facts that change faster than others. The term half-life comes from physics, where the half-life of an element is the time it takes for half the mass to decay to another type of element. Similarly, we can estimate the time it takes for half the knowledge in an area to be superseded or overturned. In fields like Medicine and others, this involves leadership change.

If facts change all the time, shouldn’t opinions follow?

Value investors have marked a very long period of under-performance. I’ve been chronicling the academic take on this for my readers.

Roger Murray Lecture

Accordingly, I looked forward to the 1st Roger Murray Lecture organised by the Heilbrunn Center for Graham & Dodd Investing. I’m always grateful to the Center for selecting me to join the Applied Value Investing program in 2010. Thus, I was looking forward to hearing Prof Greenwald speak again. I had no access to the post lecture recording and this post is based on my top of head recollections.  

The biggest takeaway was the upcoming release of the 2nd edition of Prof Greenwald’s classic, Value Investing: From Graham to Buffett and Beyond (Amazon is accepting pre-orders now https://www.amazon.com/Value-Investing-Graham-Buffett-Finance/dp/0470116730/). This has taken 17 years. It’s also timely since value has suffered a long period of under-performance. The book’s 2nd Edition needs to capture a lot of new facts that have changed since the first edition of 2004.

Value’s under-performance

Prof Greenwald opened the lecture by addressing the elephant in the room. Similar points have been raised in other fora as well.

  1. We’re in the late stage of a very long economic expansion. Value tends to do poorly at this stage;
  2. Competition is intense due to Warren Buffett, Joel Greenblatt and others popularizing value investing. It’s harder to find undiscovered stocks;
  3. Modern day growth stocks can maintain their high returns on capital as they grow. Old manufacturing /natural resource type businesses saw returns on capital diminish once they grew large enough to compete globally. I learned a new term here “Fade Rate”. This term refers to how fast the rate of return drops.
  4. Intangible capital has become more important for security analysis of services companies. Traditional measures like price to book and price to earnings become undermined unless intangible capital is captured.  We totally agree on this observation. In our view, there are two types of intangibles. Firstly, some are capitalized and reflected on balance sheets. Others are not, meaning they have been expensed. An example of the latter is brand advertising spend. Both forms require effort to estimate and adjust in order to arrive at some semblance of an intrinsic value for the stock.

He then moved on to his favorite topic of monopolies and hyper local markets. He also espoused on today’s tech companies like Google and Microsoft.

DCF…. Not

Prof Greenwald also addressed the challenge of valuing growth stocks. Prof Greenwald argues against DCF. Most of the error in DCF estimates (10 year growth rates and margins for example) are embedded in the Terminal Value. This is like adding bad information (the Terminal Value) to good information (say forecasted cash flow for next 3 years).

Bad information + Good information = Bad forecast.

His preferred method is the investment return method. You can work off relatively accurate information that you have in hand and thus end up with a reliable forecast. First, we estimate the source of returns from holding the stock. There’s a cash return (dividend), share repurchase/issue rate (net), return on growth. If there’s a franchise, then the return on growth is positive otherwise, we end up with a negative return! This is the most critical analysis for any growth company. Thus, we focus on identifying the existence of a franchise in 8VantEdge’s stocks.

Precision or Accuracy

Arbesman touches on precision and accuracy in Chapter 8 of his book. Precision is how close a repeated measurement is to each other. Accuracy is how close the measurement is to the accepted or correct value. Greenwald’s point is that the rate of return method is likely to be closer to a correct value than the DCF. Clearly, the precision of either method depends only on the number of decimal places used by the modeler.

Changing your opinion when the facts change

Prof Greenwald confirmed that he’s no longer short Amazon. No surprises there! But he wouldn’t buy it at today’s prices. Prof Greenwald was happy to share a few facts that changed his opinion. He highlighted the facts that were surprises to him such as the stickiness of Amazon’s customer base and Amazon’s ability to enter ancillary services like cloud. These facts were relevant to change his opinion. Changing your opinion when the facts change is a struggle for all of us.

Closing

We’re coming to the end of the 2nd month of the fund. 5 of our positions are negative (slightly). Everything else is up. It’s a rare situation to be in for a value investor.

In the short term, markets are positively serially correlated (i.e what goes up keeps going up). In the long term, markets are negatively serially correlated (i.e. what goes up must come down). We’re highly sensitive to this fact. The long term can be really long though.

Quitting our stable jobs and launching a new business during a global pandemic has turned out to be a blessing in disguise. Better to be lucky than smart.  

Need help?