Sunday, April 13, 2014

Working Thoughts on Long-Term Equity Returns

A few semi-random thoughts prompted by the abundance of "statistics" regarding the "stock market." I use " " because sometimes the term is very ill-defined. 

You've all seen them:
- Since 18XX/19XX, the "stock market" has returned X%/Y% per year
- Since 18XX/19XX, the LTM/NTM P/E of the "stock market" is X%/Y% per year
- Since 18XX/19XX , the "stock market" cap to GDP is XX%/YY%
- Since 18XX/19XX, the "stock market" "profit margins" are XX%/YY%

These stats are of limited use because the "stock market" of 18XX is not the stock market of 19XX which itself is not the stock market of 20XX. The stats of the 1991-93 Chicago Bulls stats are irrelevant to the Chicago Bulls today, and some of that is applicable to the "stock market". 

Some of the obvious differences are:
- Stock market composition: just as an example, railroads were a much bigger percentage a hundred years ago than they are now. 

- National GDP has a drastically different profile and is less relevant due to business globalization (and GDP calculations themselves change over time)

However, the biggest difference is that...

- Stocks, as financial assets, are MUCH closer to cash therefore their returns will be MUCH closer to cash longer term

Let me elaborate on how stock are closer to cash vs. 20-50-100 years ago:
- Liquidity when measured in share volumes: easier to transform any one stock in cash within a second for many market participants 
- Liquidity when measure in how the stock can be transformed to cash: sale, sale of covered calls, short-sale of index against a portfolio, etc.
- Liquidity when measured in transaction costs, both commissions (zero in some cases) and slippage 
- Lower diversification cost: one can get a Vanguard index fund at 0.10% annual cost
- The low cost to diversify has led to more diversification (index products), and these are the most liquid names
- Lower information risk: easier to see if a company is a fraud (incl. whether it actually exists), easier to see who the other holders are, easier to see reliable, periodic, audited financial statements, etc. 
- In some cases, partial or full sales data for companies is available in near-real time (i.e. scanner data for CPG from Nielsen)
- In other cases, company returns are pre-set (i.e. regulated utilities that are allowed to mark-up your power bill as they see fit to get the commission-allowed ROE) 
- In other cases, revenues are reasonably foreseeable (i.e. "pre-sold" or hedged production in oil and gas) 
- In general, if you think about it, there are a lot fewer "dark corners" and "black boxes" now. Most of the outright fraud is confined to the OTC markets or in regulatory capture (which benefits shareholders for the most part)

That is not to say that we won't have 1987 or 2002 or 2008 ever again: stock are still influenced by panics, excessively high or low expectations, supply and demand, margin debt/forced selling, taxation of gains and dividends, etc. as they always have been. Maybe HFT and algos are a bigger risk. But one probably should not expect the 10- or 20- or 30- year returns from a century ago when a lot more risks and costs had to be priced in.