What will happen eventually?

The world is full of point estimates and time bound estimates. Will Tesla shares (or stock market in general) fall from here over next year? Will a full rebound of the economy be possible by 2021? What will be the GDP growth next quarter? Will China be the dominant world power by 2050? The issue with these estimates in complex systems, no matter how well thought, is that the inherent randomness of the economy or market means that even the smallest of factors can influence it in a way that nobody could predict. Thus we often end up with the statement "The estimates were broadly correct, but for unforeseen xyz factor". Well there are hundreds of unforeseen factors that can have temporary influence over the estimates at least as far as economy and markets are concerned. I think this conundrum should be obvious to most people in business of forecasting.

Still millions of people keep on spending significant time and energy in making these estimates. How much of our time is spent in thinking "What is likely to happen in next 6 months or in next 1 year" even though it ends up being a fruitless endeavor in most cases. One can argue it even has a huge negative impact going by the anxieties and subsequent irrational actions it produces.

The other approach is to look back in history to estimate what will happen. So people come up with theories like value stocks generate higher returns or small caps generate higher returns or quality at any price strategy generate higher returns (based on which decade and which stock market you are evaluating). The problem with this approach is also that the patterns keep on varying across decades. So even if your historical analysis may work for some number of years, there is always a risk that the pattern will shift and you might end up losing money compared to buying the index. And when a strategy succeeds the capital invested in it keeps increasing. So when it starts failing, the capital at risk is much more compared to the time when it was succeeding. Thus eventually such strategies which focus on historic patterns and not a cause-effect cycle, end up losing money (in comparative sense) despite short term success.

In sharp contrast to these two approaches, one of the striking thought process that I have come across in writings Warren Buffett is his tendency to identify continuing cycle with compounding effects rather than trying to make time bound estimates or try to find out historic patterns. If a cycle has a predictable outcome, it continues for a long time and has compounding effect you can easily bet on the eventual outcome and can escape the tyranny of fickle factors with temporary effect. Thus this approach helps him to predict what will happen eventually even though the interim path may not be clear. Buffett's approach mostly is about benefitting from such cycles while making sure that a short term random factor doesn't spoil the party in the interim.

For example, in one of his early articles, he has described companies as "cash-in / cash-out" cycle (or what I call as ROE engine). The system can be described as follows: A company invests money in fixed assets and working capital which produces revenue. A portion of this revenue is spent in fixed and variable costs, interest (if the company is leveraged), replacement capex (to take care of wear and tear/reduced useful life of fixed assets), and tax. The resultant cash flow is your return (assuming you had invested in the company when it's market price was equal to its book equity). But the engine doesn't stop here. Most good companies can simply reinvest this capital back in the business because there are good growth opportunities available. Furthermore, many of these companies can get even higher return in future from these investments (things like economies of scale, network effect, pricing power etc. means that lesser capital needs to be invested in future for more returns). If the company operates in a stable market, has sustainable competitive advantage, stable business economics, and a prudent management you can be sure that the company can keep on producing higher cashflows and returns in future. In short you get a continuing cycle with compounding effects.

But no matter how good the company is, after a few decades the market stagnates and the competitive advantage erodes. Instead of becoming an advantage, the legacy become a shackle to further growth. In such a situation, what can be the possible method to keep the magical cash in - higher cash out process keep running for a longer time? Simply overlay the same process to multiple companies. What do you do to achieve this? Make sure that the companies you have originally invested are good at capital allocation. This means that they invest in their own business or similar business if they are sure of high rate of returns, and if they cannot do so then they return the excess cash to investors (after making sure they have enough cash for a rainy day). Thus as an investor you can accumulate enough cash from your investments and invest them in newer ventures. Just as a company can re-invest it's cashflows to it's own or similar business, you re-invest your cash proceeds to new ventures when you are sure of high rate of returns (i.e. stable business with high ROE and reasonable price). If done right, this again results in a even longer cash in - higher cash out cycle with compounding effect.

This is exactly what Berkshire currently does. When people try to evaluate Berkshire by it's components they forget it is more than that. It also consists of great companies that it can acquire in future at reasonable price. Even in the past, their top holdings of Berkshire have differed every decade. There can be years or even decades when they cannot buy enough new businesses because of continued high valuations, but eventually the opportunity to buy great businesses at reasonable cost will present itself.

Individual investors can also learn from this process by properly understanding the process of compounding. Compounding cannot be viewed as %age price appreciation you can expect from a particular stock or from the growth rate of the company. Compounding results from stability of the business, growth opportunities the business has, marginal return on equity and disciplined capital allocation. To benefit from this cycle of compounding, instead of doing point optimizations, an investor has to do process optimization. It is impossible to tell at all times whether a stock is overvalued or undervalued and whether the share price will go up or down. But it is possible to tell that a process that involves buying good companies which are good at capital allocation at the reasonable price and a process that views cash as an option to buy into good businesses at reasonable price in future will succeed eventually.

This is just an example of how betting on compounding cycles and watching out for things that can kill you can lead to massive returns in future. But I believe it is more of a general principle. The same principle can be applied to many different fields. More on this later.