What Drives Stock Prices?

Although likely considered basic and foundational for seasoned investors, it is mandatory reading; full stop.

"I’ve said forever that people always ask me about valuation of individual enterprises, when half of my career and time is spent thinking about how those pieces fit together in a given portfolio. Hardly anyone asks about the latter, yet it matters enormously. I think about weightings a great deal."

I guess I keep asking the wrong people because hardly anyone will give me a full/thorough answer. Either they don't know, or this is the secret sauce that they are unwilling to part with. How exactly should someone determine appropriate portfolio parameters and then implement investments to match the parameters? To get a bit more technical, how would someone determine their risk tolerance and capacity (Are these "feelings?" Are they numbers? Can they be calculated?). Related to that, how can they determine maximum tolerable drawdown, tolerable standard deviation, etc. What is the full set of parameters that should be considered? Once those parameters are determined, how to you select a portfolio that aligns? I do understand the basics of modern portfolio theory and the efficient frontier, but how do you actually implement that - or whatever framework you use - in practice?

Thanks!

So, keeping this academic and high-level educational (like all things on my personal blog, none of this is investment advice), it might help to think of it like this ... and please forgive any typos or de minimis errors because I am trying to bang this out over a cup of coffee while Aaron takes the kids to school before we head to the office. I'm typing it in real-time and won't re-read it before I submit it. Think of it like us having a conversation; spoken word.

First, you have to take a step back. Foundational. Every pool of capital, or portfolio, has a job to do. Many successful families have multiple portfolios. You start by clearly defining the goal of that portfolio.

To provide a few illustrations:

At the investment firm I own, we have several different types of clients.

One group consists of highly successful individuals, some of whom even work in finance, and aggressively compound their money. They shift a portion of their capital to a portfolio under our management, typically in the seven or eight figures, and have us manage it on a Global Value basis. This is inherently more conservative than their main activity, or portfolio, that they manage which can consist of taking over entire businesses, acquiring or developing real estate, launching technology startups, or working on Wall Street. The capital we handle for them is designed as a sort of "permanent capital" with the view that if it fails, civilization has fallen for all intents and purposes. Some dump in a large lump sum and plan on letting it ride for decades, others regularly add five or six figures with the goal of expanding it over time and re-arming us with dry powder for opportunities, but the point is, if God forbid they died, the complex, riskier operations they oversaw gets sold by their estate and the bulk of the rest of the money ends up at the Firm alongside their existing holdings, allowing the surviving spouse or children to maintain a high standard of living.

Others have unique twists in the mandate, such as tech executives earning millions of dollars a year who request we build a portfolio using the same philosophy but avoid all technology companies regardless of what that may do to performance because the job of the pool of capital is to provide a ballast of sorts for human capital + equity rewards that are entirely concentrated in a single industry that, at some point, will likely experience another horrific winter like the 2000-2003 period.

Still others are retired, or experienced a liquidity event, and want to live off their money as a source of private income now.

An insurance company might have a book of annuities that it needs to fund, which involves an asset / liability matching component.

A deeply religious investor might want to earn money but not worry about holding certain types of companies.

A pension fund manager for a small business may need to think about trying to control returns within a statistical band (as strange as it seems, for something like a Cash Balance Plan or similar, if it isn't designed in a specific way, having performance that is too high can cause some horrendous problems so you want to target a 5% to 7% range long-term return; way beyond the scope of this discussion but it gets into end date ultimate benefits versus the potential for confiscatory excise taxes).

All different jobs.

So, again, before you look at construction, and the weightings that arise from that, you first have to be clear about the job. Most small investors without a lot of capital think the goal is to compound money at the highest rate possible, timing, taxes, cash flow, and risk be damned. Once you have any real money, that is not at all how your life operates; college funds that need to be there at a specific time even if we are going through a 2009; purposely taking losses on positions you otherwise want to hold to maximize the tax-vs-time-value-of-money calculation with the intention or hope of repurchasing in the future (esp. for high earnings in states like New York, New Jersey, or California); strategic liquidity reserves that should be available within six to twelve months at all times to support a portfolio of privately controlled businesses, etc.

Once you identify the job, you then start to work backwards through a series of checklists. First, is there an asset/liability matching component. That is probably the most important thing to settle from the beginning. For example, if someone shows up at the Firm with, say, a $2,000,000 portfolio and wants to retire in the near future, taking a distribution of, say, $6,000 per month, that money has to be non-negotiable to extent possible in the capital markets even if we hit a 1929-1933 tomorrow, which was the worst economic event in 600 years. That is a clearly-defined asset/liability matching problem. To survive a collapse like that, you need anywhere from 36 to 60 months of reserves held in the highest quality asset possible consistent with doing that job, which for U.S. investors certainly is a U.S. Treasury backed by the taxing and military power of the United States, compelled to be paid by the U.S. Constitution, and exempt from state (but not Federal) taxation. We treat the $500,000 or $2,000,000 client the same way a family office would treat a $70,000,000 client (because that is precisely how I expect my own capital to be treated and my name is on the building) in that I will literally build out something like a 60-rung or greater maturity ladder of U.S. Treasurys the client owns themselves, in a segregated portfolio. It's far more powerful than, say, a 5-year ETF where the maturity is constantly rolling because you have specific contractual cash flows guaranteed by the government on specific days with clearly defined coupons, yields-to-maturity, and risks. So in the case of our hypothetical $2,000,000 investor, you're, by definition, talking about $6,000 x 36 = $216,000 (10.8% of capital) or, much more likely at the start, $6,000 x 60 = $360,000 (18.0% of capital) for what I call the "firewall" component. I use firewall because if the world is burning down, this is the thing that has given the client time. If stocks are down 90%, unemployment is 25%, there is another Dust Bowl, etc. those distributions should continue as long as the United States Government has not fallen. (If it has, so too has the global economy given its size and the interconnected nature; no one of note or scale escaped.) Importantly, those Treasurys are the same type of asset that companies like Berkshire Hathaway hold when you hear about their legendary "cash" reserves. It's largely not cash, it's directly identifiable Treasury securities.

The point is, that by defining the withdrawal rate (e.g., $6,000 per month), and applying a 60-month reserve based on the 1929-1933 experience (which would likely last longer because the typical person is going to severely reduce their spending in the event of such an economic cataclysm) has set the first group component weighting (firewall at 18.0% of $2,000,000 capital, each rung at $6,000), but if the withdrawal rate were the same and the portfolio had started at $5,000,000, the same "job" took only 7.20% of capital. You start with the job. You achieve the job. It definitionally absorbs some of the capital.

(Now, in actual practice, I may build in inflation increases for some clients under some conditions, and I will also accept "brokered" FDIC-backed certificates of deposit provided the client's aggregate exposure to any one institution is below the insured limits because many will have special features such as a spousal "Survivor's Put" so that if one spouse dies, we have the option to demand full par back as an emergency, which could be powerful under a set of horrible circumstances you hope never happens. Again, I'm simplifying this to high level principles.)

Beyond that, if the person is living off the capital presently, and they hold no other outside investments, I'd typically want the sum of [cash + the firewall + general fixed income] be no less than 25%, and no more than 75%, of the portfolio in question (and remember many successful families run multiple portfolios, some at a single institution, often at many institutions) given my objective is to have them survive both a 1929-1933 scenario and a 1970s-style inflation run. There are no guarantees, of course, but mathematically / probabilistically, it's what lets me sleep at night. This allows you to adjust for valuation levels; e.g., if it's the late 1990s and Coca-Cola is trading at 50x earnings while U.S. Treasurys are yielding 6%, you obviously don't want a lot of stock, whereas if it is 2009 and everything is on sale, especially as a value investor, you probably want to be buying businesses left and right. Again, we definitionally know that the withdrawal rate is what set the absolute weighting of the firewall component, so you look at age, life expectancy, other plans (you may have sub-portfolios within this set of components to fund specific things like construction of a vacation house, etc.) The general fixed-income component can consist of anything subject to ordinary risk, valuation, etc. There has been a brief moment in recent weeks where, using an asset placement strategy, I've been able to build up fixed-income components in tax-sheltered retirement accounts locking in roughly 5.5% to 6% yields-to-maturity and yields-to-call holding investment grade positions in solid underlying companies. Other times, I've bought plain-vanilla mortgage backed securities with great underlying characteristics so you see the principal and interest come in each month. Some involve tax-advantaged municipal bonds. It's a combination of what is attractive, where the risks in the economy are, what the yield curve looks like; etc. Individually, I think a lot about overall weighted duration, sector/industry exposure; you get the idea.

After you've solved these issues, you're left with equity weight (or if there were no fixed-income component, you start here in the case of a 100% equity allocation). Importantly, you think about not just absolute weight of a specific component, but relative to other components. For example, a great investment, weighted at 5%-at-cost or, in some cases, 6.5%-at-cost, can really move the needle over a decade or two. Importantly, if you have another non-correlated holding weighted at the same thing, it can overcome a lot of disaster through what I used to call "The Home Depot effect" ... e.g. if you buy one at 5% and it returns a 500% profit, it mathematically swamps a second component that goes down 50% and is sold at a loss. Consider a $1,000,000 equity allocation with $50,000-at-cost put in Holding 1 and $50,000-at-cost put in Holding 2. If, seven or ten years later, Holding 1 is $250,000, it doesn't matter much if Holding 2 is at $25,000. In an ideal world you want to pattern-match, for lack of a better term, so that Holding 1 and Holding 2 are not the same. Ideally, you want to do this on an industry or sector basis, too, but what really counts are individual holdings at specific prices. For example, I'm not going to wake up and say, "I need to put [x%] into utilities or banks", I wake up and say, "Let me study everything going on in the world and see what is interesting on a risk-adjusted basis where I think the probabilities over the next decade or two are strongly in our favor even if things go down 50% the day after I buy them."

Sometimes the nature of how the cash flows are delivered can change things dramatically. For example, I responded to another comment last night using a real-world example of something that happened early on in the decade. I won't repeat the entire thing here ( it is on the Lessons from Ray Kroc's Paper Cup Years post ), but the portion relevant to this conversation read:

"Sometimes the nature of how the cash flows are delivered can change things dramatically. For example, I responded to another comment last night using a real-world example of something that happened early on in the decade.

That's not to say every apparently obvious situation works out. There have been a handful of disappointments, but the thing about compounding is the math works out wonderfully if you think about portfolio structure."

In some cases, especially for wealthy folks in their 20s-50s with all equity portfolios and who were regularly contributing new deposits, I built out the group operation at about 1/3rd of the equity capital split among those components; one structural trade representing a third of money with some individual components going to I believe it was 8.5% but I'd have to check. The point is, I never would have done that if the companies had a 100% retention policy rather than sending most of the cash out the door in the form of dividends. Buying at such high yields-on-cost when the payouts weren't in question - these businesses were not distressed by any stretch of the imagination - the relative risk dropped with each passing quarter as we were drowned in cash flow. Like clockwork, another 90-some days pass and a wave of deposits hit across the accounts, fueling acquisitions in other non-tobacco holdings. The stocks went nowhere for something like 3 years - investors simply did not want them - until suddenly they did. It didn't matter. We were holding them to extract the cash, like an oil well that you know might eventually run out of oil but the initial outlay was so attractive it wouldn't matter if it did. If, in 20 years, those stakes all went to zero, the overall wealth creation was still meaningful, especially relative to alternatives, risk, and taxes. So, point being, if they had been zero-dividend paying companies with similar profiles, rather than going to 1/3rd of capital at the time, I might have limited it to 1/5th or something to account for the fact I couldn't take money off the table without selling shares or writing covered calls.

I have to go ... this is a huge, huge topic. There is no way I could ever condense it to a single comment or even a series of blog posts. It's is as expansive as things I've written about on valuation, case studies, etc. over the years. Again, none of this is investment advice, I'm trying to explain it like I would if I were a professor going over the 100,000-foot view of an area of study. To that extent, I hope at least something in it was helpful or educational.

P.S. I really do have to run but, in summary, to answer your first question, weightings at the first level are determined by asset/liability matching. That's principle one. The other questions, again, are each their own 3+ hour conversation. I mean there are two or three models you can use for maximum draw-down; a fixed-percentage will be informed by time-sequence risks with equities, which the firewall component I mentioned as part of asset/liability matching mitigates to a large degree, in my opinion, typically no more than 4% but there is a lot of potential adjustment in that number, life expectancy, etc. Usually, if people want to travel, you want to have them do it before they are 70 because at that point, no matter what the intention is, most people become more stationary and stay at home for one reason or another. Not everyone, but enough the presumption should be factored into the math I'd argue; better to take 6% for 5 years then maybe 3% down the road thereafter if it means you get so see Italy, France, Greece, Japan, or whatever your particular list may be.

You just go through these massive mental checklists and then, when all is said and done, I stand back and sit with a portfolio for hours, thinking about the person, what they've shared about their life, the specific cash flows at the underlying components, the weightings in relations to each other, and mentally run through horrific scenarios; 1929-1933, the stock market closing for a year like in World War I, another 9/11 event, etc. I basically ask the question, "If this were my portfolio, and I have the same facts, preferences, and consideration as this client but with all of my knowledge and experience, would I sleep at night?" The answer has to be yes or everything gets shut down until I get it to a place where I feel that comfort.

So there is a tremendous amount of math baked into it but by the time I'm doing the work, and adjusting weightings, it's second nature. I also then look at the equity weightings and compare it apples-to-apples with different lists like the S&P 500. Right now, for example, the S&P is essentially one major concentrated bet in one major industry, the ultimate cash flows of which look really iffy. It's certainly the most concentrated and speculative it's been in my lifetime, or at least since the dot-com era, while other businesses are really, really attractive at present prices. Hence, our capital looks very different than the headline market number with entirely different weightings; e.g., the tobacco structural trade is a double-digit holding to this day for most folks because even though portfolios have grown, they have expanded, too, and are far above the initial cost basis. Yes, there is some overlap - e.g., we have several major tech positions, Meta and Alphabet plus a third unnamed one that is now becoming quite large, bought during brief moments of collapse in the shares because investors thought the companies were doomed and we believed strongly otherwise - but there may be periods where the S&P significantly outperforms us, or we significantly outperform it, I simply do not care to the extent that some of the things happening in the index right now are insane. I would be terrified if my entire net worth consisted of, say, a pile of index funds in a taxable brokerage account. I've seen this movie.

It reminds me of something Charlie Munger once said, you go around trying to make intelligent decisions and you deviate wildly from society or the market, and sometimes you can be out of favor for quite some time but then, and my own life experience has mimicked this, something happens, and all at once you're in the right place, with the right holdings, having avoided the nonsense that infected others. I said it elsewhere in recent days, but my goodness, I'd be terrified if I worked at OpenAI or Anthropic, counting down the days to an IPO because I'd want to take the liquidity as quickly as possible. No matter how many times I run the numbers, I cannot get to any other conclusion that LLMs are mostly going to end up being locally run on company-owned or consumer-owned hardware with little to no moat in the frontier models. The benefit should accrue to the Alphabets and Apples and Microsofts of the world who, over time, just bake it into existing platforms. Other than specialty product (e.g., a company like Thomson Reuters creating a legal LLM model that it provides to law firms and couples with a specific indemnification guarantee), I think paying for tokens with some of these companies is going to be akin to how internet browsers used to function. No one has bought an internet browser, largely, since Microsoft made Internet Explorer free and bundled with Windows, destroying Netscape's business.

And the thing is, all those early internet folks were absolutely right. Their predictions came true. More so, even. They simply lost most if not all of their money in the end and existing, large businesses integrated the improvements; e.g., the internet did not destroy Walmart, it has now made it even more unstoppable. (Valuation on Walmart these days is another question, we're talking about underlying cash flow growth and operational stability.)