investing

Thoughts on Inflation

I am far from a macroeconomics expert. But I do think that macroeconomics provides a valuable perspective on the world, so occasionally I like to take a step back and try understand some of the macroeconomic dynamics at play.

One topic that’s top of mind for many folks today is inflation. There are two common comments I hear about inflation these days. One is that – prior to the past few months – inflation has been surprisingly low for an extended period of time. Between the government bailouts, the quantitative easing of the Fed, and the increasing amounts of debt in the economy, many folks expected the rate of inflation to be much higher than what has actually occurred.

The second comment is that now things are finally going to change. With the further massive stimulus during COVID, the beginning of the recovery therefrom, and potentially two additional massive infrastructure bills, the risk of runaway inflation is at an all time high.

Past…

Let’s start with the first comment. Up until very recently, why haven’t we seen much inflation?

Inflation occurs when prices go up. So let’s look at prices. Below is an old chart from the American Enterprise Institute from 2020:

The first thing to notice is that the amount of inflation strongly depends upon what you’re looking at. Generally speaking, things that have benefited significantly from technology – either because they can be manufactured more efficiently (i.e. TVs) or can scale at near marginal cost (i.e. software, Netflix, etc.) – or from globalization have actually gone down in price, while things that are still require local human labor continue to increase in price. (An interesting exception on the chart is college textbooks). The second thing to notice is that, in aggregate, there still has been inflation of around 2.2% annually on average. So not only has there been inflation, but it varies widely by category. Measures like the Consumer Price Index can mask this.

To a first order approximation, prices are related to the ratio of supply and demand. As I mentioned above, the marginal cost to supply an additional digital good is virtually zero. I don’t have the data, but I’d also assume that there has also been a significant shift by both consumers and businesses to spend more money on software and digital services.

This could partially explain the relatively low rate of inflation. One the one hand, a product or service with nearly infinite supply (software, streaming video, videogames, etc.) can accommodate nearly any increase in demand without triggering a price increase. On the other hand, any money spent on these things is money not spent on things that do have a marginal cost to supply; this controls demand (and thus prices) for those items.

...is Prologue

This explains why we haven’t seen that much inflation over the past two decades. What about moving forward? We’ve recently seen a significant uptick in inflation – surely all of the public and private debt-fueled spending and quantitative easing is catching up with us, right? Isn’t this just the beginning of massive hyperinflation?

I’m not so sure. It is true that we’ve seen a recent spike in inflation. But this seems to me like it may be short lived.

Here’s my reasoning: when COVID hit, there were a few effects that conspired to increase prices. On the demand side, because people were locked down, demand for services decreased while demand for durable goods (TVs, housing, etc.) increased.

At the exact same time, due both to some people not working and significant changes in demand for some items (like masks and other PPE), supply chains faced significant disruptions. This limited the supply of these exact same durable goods. Higher demand and lower supply equals higher prices. So we see inflation.

However, it seems to me that these same dynamics could soon lead to the opposite effect. Because people “pulled forward” their demand for things like TVs, exercise equipment, and home renovations, there is likely to be weakened demand for those things in the mid-term. Simultaneously, as supply chains get unjammed, foreign goods are likely to flood the market with additional supply. More supply and less demand? Leads to lower prices (or at least stagnating ones).

And there’s more. During Covid, employers also “pulled forward” investments in technology to support remote work. These investments are improving productivity. It is true that labor shortages in some areas are creating wage increases. However, if productivity increases faster than wages and other costs remain steady, then the the per-unit cost of production goes down. (e.g. if I have to pay you $60/hr instead of $50/hr but the number of TVs I can produce an hour goes from 100 to 500, my cost-per-TV has still gone down).

As foreign goods comes back into the market, domestic producers will want to try to keep the market share they obtained during the pandemic. Whereas prior to the pandemic they might have been beat out on cost, because of the reduction in their unit cost of production described above, they will have more ability to compete on cost than before, further driving down the price of goods.

Blue: Inflationary Dynamics. Brown: 2nd-Order Deflationary Ones.

Conclusion

I thus think it’s reasonable to expect that we see a significant reduction in inflation – or perhaps even mild deflation after this transitory inflation spike. We have a situation where technology is reducing the unit cost of production of goods while likely reducing the demand for labor (thus lowering wages and thus consumer spending), all set against a backdrop of significant debt. Unless that debt is invested in such a way as to produce cashflows adequate to service that debt and reinvest (which, in aggregate, I doubt), then it simply pulls forward demand to the present, thus limiting demand in the future. Lower demand and cheaper, more abundant supply leads to lower prices and disinflation.

While there are scenarios that could potentially lead to hyperinflation (e.g. the Fed being given authority to spend or a decision to intentionally debase the currency to reduce debts), I think the more likely scenario is transitory inflation followed by disinflation or even deflation. From a portfolio perspective, this would suggest positioning oneself for an expectation of mid-term (dis/de)flation (e.g. cash, defensive equities, long-duration treasuries) with a hedge to try to protect against extreme hyper inflation (gold, real estate, crypto, equity puts, etc.)

As I mentioned at the outset, I’m no expert in this space. But this is my current thinking. Let me know where I may be wrong.

What Causal Inference Can Tell Us About Hiring

One area I’ve gotten interested in lately is causal inference. For those of you not familiar, it’s a methodology that attempts to find and validate cause-effect relationships between variables. The key is that it attempts to do so using data without having to rely on controlled experiments. (For an introduction for the casual reader, I highly recommend Judea Pearl’s book The Book of Why.)

One concept I found interesting was the implications of something called a collider. A collider is a variable that is the effect of two or more variables. As a simple example, consider the following:

The way to read this diagram is that fame is a function (or effect) of money, talent, and looks. In other words, fame = f(money, talent, looks). In this example, fame is a collider relative to money, looks, and talent because they all have arrows pointing into fame.

The interesting implication from the book is the following: given that you hold the level of a collider constant, the other variables become dependent upon each other even though there is no causal influence on them.

To understand this better, let’s use an even simpler example: X + Y = Z. In this case, Z is a function of X and Y (i.e. Z = f(X,Y)), so Z is a collider with respect to X and Y:

Here’s the key point: if we fix the value of Z at some specific value (say 10, so we’re left with the relationship X + Y = 10), then X and Y become correlated. In other words, if I know the value of X (say 8), then I can infer Y (i.e. 2).

The interesting finding from causal inference is that this dynamic generalizes. Said another way, for a given level of Z, information about X automatically gives me some information about Y, even if I can’t observe Y directly.

Almost Famous

Why is this interesting? Let’s go back to our fame example. Assuming our causal model is valid, then we can say that for a given level level of fame, if we know something about their level of wealth, we can infer something about their level of looks and talent. If we simplified it down for a minute to say just include looks and talent, then we could say – for a given level of fame – we’d expect that a person who is more attractive is likely to be less talented. (Another way to think about this is if they were both attractive and talented, they’d be even more famous).

I haven’t done an analysis to verify this yet, but it’d be interesting to run an experiment. For example, look on social media for actors who have a similar level of followers (as a proxy for fame). Within that cohort, if the model is valid then you would see a spectrum ranging from the good-looking-but-hacky to the talented-but-ugly.

Counterintuitive Hiring

This finding has interesting implications in many places. Take hiring, for example. Consider, for example, a hypothesis that seniority_level = f(skill, likability). If you think both skill and likability are positively correlated to seniority level, then – for a given level of seniority – consider that the most skilled person is likely to be the one you personally like the least.

These are of course toy examples; the causal structure of real life is likely to be much more complex. But they illustrate both the power of causal analysis and the sometimes counterintuitive truths behind the way the world works.

Investment Dissection: Homebase.ai

[Note: My posts on this blog vary in length. The last one was long. So is this one. The next few will be shorter.]

I recently made an investment in Homebase.ai, a “smart apartment” platform. In this post I’ll walk through how I thought through that decision. I’m doing this for two reasons. First, I think it’s good practice to keep a ‘decision journal’ that one can review and learn from. Second, I’m hoping that others can point out things I missed or where I could be wrong. For confidentiality reasons there are many data points that I can’t share. But that’s okay because I want the focus here to be on the process.

Here are roughly the steps I went through, along with how my thinking on how the opportunity evolved over time.

Step 01: Intro

What Happened
I heard about the opportunity from a friend who had invested at an earlier stage. The company was now raising money at a significant increase in valuation (roughly a 5x increase in ~18 mo) and had seemingly good traction. There were also two other similar companies that had recently gone public via a SPAC: Latch at a ~$1.5 billion valuation and Smart Rent at $2.2 billion. Homebase was much smaller and earlier stage.

My Thought Process
(First, an aside: when evaluating investments I think of them in terms of probability distributions. The shape of that probability distribution adjusts as I get more information. As I talk through my thought process here, I’ll reference this payoff distribution frequently. OK, let’s get back to it.)

At this time the SPAC market was still quite hot. While my general view was that the SPAC market was overheated and valuations were getting inflated, I have learned to both develop my own independent opinion of a company’s worth and to understand that – when actually seeking liquidity – the price of an asset is what the highest bidder is willing to pay. I was also aware that while the hot SPAC market could disappear at any time – it hadn’t yet.

Moreover, while this company was still early stage, the fact that two competitors had recently gone public told me a few things. First, at a minimum there were clear, public comps available. This would allow me see how Homebase compared from a metric-to-valuation perspective. Second, given the general SPAC frenzy, I thought that this might make it easier for Homebase to go public via a SPAC in the future since the approach had been ‘proven’. Third, I figured that even if Homebase didn’t go public that way, there were now at least two competitors in the field with a lot of cash that could potentially acquire them down the line. Finally, I thought that the fact the fact that two companies in this space had gone public might indicate that there was something fundamentally successful about the business-model / timing interaction (i.e. it was a sign of a fundamentally good business model at the right time in the market and not just some rockstar team or fluke of luck). Given how hot the SPAC market was, however, it was also possible that this was just another company being sold on hype (I’m looking at you, WeWork).

For all the reasons above, this deal immediately caught my attention enough that I decided to investigate further. Given that the company was already beyond a typical “seed” or “pre-seed” valuation, in my view a ‘spray and pray’ 1/N approach wasn’t appropriate. Therefore, I knew that to even invest at all I needed to feel comfortable that there was a real business there at a reasonable valuation. However, I also thought that the SPAC route might provide a potential far right tail on the distribution, which might effect how much I would invest. I’ll talk more about this later.

So I started digging.

Step 02: High Level Market and Competitor Analysis

What I Did
Though I wouldn’t always do this step first, because of the two recent SPACs, there was a treasure trove of relevant data that I could find just by looking at the public documents that each competitor had provided to investors during the process (which were available on their websites), but it was also relatively easy to find other people’s analysis of those companies. I could then triangulate at least some of those data points with other sources just to make sure the numbers seemed reasonable.

I first started with the general market and my knowledge about it. Homebase’s general market was the “rental home / apartment space” since they are (mostly) selling to contractors building new buildings and landlords. I knew that real estate generally was the largest asset class in the world so that part was fine, and intuitively I felt comfortable that the size of the rental market was pretty large. A quick search told me that there are about 47M rental homes in the US and 93M apartments in Europe. Without nitpicking, that seemed ‘big enough’ to me. Now I didn’t yet have a feel for what ‘end’ of the market (i.e. luxury, mid-market, low-income, etc.) Homebase was really targeting so I didn’t do a deeper analysis yet on that, but I did make a note to come back to that later. I also knew that real estate was likely undergoing a significant transition, which had only been accelerated by COVID: more remote work, at least a temporary flight from large cities, and likely a more permanent reduction in demand for much commercial space. That said, people still needed a place to live, so in aggregate residential real estate would likely continue to be needed. I also knew that, at least prior to the pandemic, Millennials were renting longer and many said they preferred renting over owning. (I didn’t evaluate whether or not that had shifted due to the pandemic, which I probably should have). However, I did think quantitative easing coupled with the pandemic was likely to continue to push property prices higher (at least temporarily), which would make it more difficult for folks to afford to buy even if they wanted to.

I also believed in the general trend of digitization of the home/apartment. Products like the Nest smart thermostat, Amazon Echo, Google Home, etc. continue to grow in penetration, and it seemed reasonable to me that landlords were a next segment.

In short, the general trends seemed favorable.

Next I took at look at the financials. While I can’t talk about Homebase’s financials because they are confidential, I can talk about Latch’s. Take a look at just a few (these are from their SPAC presentation and so were accurate as of that date):

If you compare those metrics to other public SaaS companies, those are near the top of the list. Importantly to me, Smart Rent had similar metrics so this said something about the general sector/model and not just about one company.

My Thought Process
There were a couple of things I took away from the above exercise. The first was that, given unit economics, unless I was missing something (which I was very aware I could be), the fundamentals of the business seemed pretty good. The second was that, while there were some network effects at play, this didn’t feel like a “winner take all” market, though because of the high switching costs (is a competitor’s electronic lock going to be so much better that you’re going to want to switch out all the locks in your building again?), I figured there would be a bit of a “land grab” dynamic, where part of the game was simply getting to scale quickly and “grabbing” as much territory as you could. In this scenario I figured the most likely outcome is an oligopoly situation.

From my research above I figured that the market was large enough that Homebase still had a chance to become one of the large players. This extended the right tail. However, I also figured that even in the situation where they become only a “medium size” player, that positioned them reasonably well to be fought over by any other bigger players with large balance sheets. Not a bad place to be. This shifted the ‘mass’ of the distribution right. The main risk then (from this perspective, anyway) was that they would grow so slowly that they wouldn’t be able to capture a footprint large enough to be valuable.

Step 03: Zoom in on the Company

What I Did
The first thing I did was speak to the team. While what I look for in teams is beyond the scope of this post, generally speaking what I’m looking for falls into two buckets: individual level things (e.g. intelligence, drive, self-awareness of a particular kind, relevant experience, etc.) and team level dynamics (is the team truly aligned on the mission, strategy, risks, their respective roles, etc.).

I again won’t go into a lot of detail here for confidentiality reasons, but I did speak to folks on the team and tried to understand their backgrounds, roles, and outlook. The reality – by their own admission – was that the management bench wasn’t that deep and, at least on paper, wouldn’t be what you’d consider a ‘dream team’ compared to their competitors. It was also true that (some) of their metrics were not quite as far along as their competitors had been at the same age. Finally, they were dependent on one particular relationship that was tremendously powerful for them but also could be problematic if that relationship went bad.

On the flip side, there were many positives. First, the reason some of the metrics were not quite as far along had to do with some specific strategy choices they made early on which would cause them to move a bit slower at the beginning but then – if you believe their thesis – would allow them to scale more quickly from then on. To me, this thesis was the crux of the differentiation and so the question became “do I believe it?” More on this later. Lastly, partially because of this strategy choice (among others), Homebase had gotten nearly as far as competitors had in the same amount of time, but with substantially less capital required.

My Thought Process
Now here’s where I might start to get a little controversial. See, I didn’t necessarily take the lack of an existing rockstar management team as a negative. In fact, this actually only reinforced the fact that this was likely a good business to be in. Why?

Warren Buffet has a famous say that goes something like this: “When a management with a reputation for brilliance tackles a business with a reputation for bad economics, it is the reputation of the business that remains intact.” I think this is a wise statement. But then I would also argue that a version of the contrapositive is also true: if you see a business with good economics with only a “good” (as opposed to great) management team, it’s probably an indication that the underlying business is a good one to be in. Think of it like a movie star. If someone is a movie star, they are usually talented, good looking, or both. In aggregate, looks + talent needs to cross some threshold. The less good looking they are, the more talented you can probably infer they are (or the son/daughter of the producer). In my view this actually decreased the weight given to the left tail of the outcome distribution.

Beyond this fact, the team readily admitted their weaknesses. Self-awareness and humility (at least of a certain kind) can be very helpful to a team if it means they will do what is necessary to supplement or replace that team to shore up gaps. This slightly shifted the probability mass to the right.

The obvious other part of this is understanding the relative valuation of Homebase vs other competitors (and other deals generally). Based upon the metrics and comps, the valuation cap and discount on the SAFE appeared to value Homebase at a reasonable discount given its earlier stage and less mature management team.

Step 03: Contraindications & Deeper Dive

What I Did
By this point in the process I was leaning towards making an investment so I flipped to the risk side. I asked myself a few questions:

  • (1) Assume things didn’t work out well – what are the most likely reasons why?
    • What evidence do I have today to suggest how (un)likely these are to occur?
    • How aware is the team of these these things and what plans to they have to reduce their probability, impact, or both?
  • (2) What are the core assumptions that must hold true for this to be a major success?
    • Do I believe they’re true?
    • Do others I consider credible in the relevant area believe they are true? If not, what is their rationale and do I agree with that?
  • (3) What evidence, if I found it, would cause me to change my mind?
    • Does that evidence exist?
    • How can I find it?

Using this as my guide, I came up with a specific list of questions for the team and asked them to provide answers and relevant information where possible.

Going through all of this in complete detail would make this post even longer than it’s already going to be, so I won’t do that here. For now I’ll focus on one of the core assumptions I spoke about previously, which had to do with a particular part of their strategy that was different from their competitors. For this particular point, I tried to drill in to understand why they thought their approach was better. I saw the logic and merits of their approach. But I could also see the risks and downsides. Ultimately, I felt I couldn’t determine which approach was better. More on this later.

My Thought Process
My goal in this process is two-fold: first, to maximize the amount of information (in the information theoretic sense) that I get while keeping the burden on the management team in check. Second, by hearing (or reading, which I personally prefer) their responses, I get a deeper insight into how the management team thinks about these issues and how deeply they’ve thought about their business.

I didn’t get all the answers I asked for, but I did get many of them. I would say that the answers I got in return were about what I expected: there were several of the questions that the team couldn’t answer well, simply because they were difficult questions and the future is uncertain. A few of their answers allayed my fears in certain areas, and a few were not as well thought through as I would have liked.

The overall picture their responses painted was mixed: on the one hand it reinforced the lack of depth on the management bench and the amount of thinking they had put into certain aspects of their business. On the other hand, part of the reason they likely hadn’t put much time into thinking through these things was that the business was growing so rapidly that they were understaffed. Net, I would say this exercise shifted my perceived probability distribution to the left slightly.

Step 05: Act Like a Customer

If you’re talking to the management team of a company, it’s probably likely that you’re going to get a relatively favorable view of the organization. Not only are their incentives aligned, but many executives are executives in part because they are good sales people. Thus, when it makes sense, I like to put myself in the position of a buyer.

What I Did
If I actually was thinking about buying Homebase’s products and services, what would I do? I’d browse the website, I’d search the web for reviews, I’d contact and talk to their salespeople…and then I’d do the same for their competitors. So that’s what I did. I found that all three companies were responsive. Latch was the most polished and professional experience of the three and Homebase the least, but Homebase made up for that with some ‘Midwestern hospitality’ (they are based in Kansas City), and the salesperson I talked to was very genuine and friendly. One thing I didn’t do, but should have, was ask each group more directly about the other to get their take; I think that would have been very helpful. I also did the mandatory Googling around to find reviews of the companies, their products, and their app. (Note that my wife and I do own real estate investments so we could evaluate as actual potential customers, I would do something similar even if I weren’t in a position to be a customer.)

In terms of Homebase’s product offering and pricing, there were actually some limitations I was saddened to hear about (namely, they just couldn’t unlock your door if you got locked out), but those limitations were imposed for legal and security reasons, which affected all competitors. The smart lock offering didn’t seem compelling enough to us alone at the moment, but when combined with their wifi solution it seemed interesting. Unfortunately, their wifi solution is much more applicable to larger buildings and/or new construction. Thus, while not a good fit for us at the moment, it was clear to me under which situations it would make sense for us to purchase, and that situation seemed reasonably likely to occur in the medium term.

My Thought Process
From this exercise, it was clear that both Latch and SmartRent were more mature, polished businesses and that Homebase was still acting more like a scrappy, lean startup. It also became clear that there was a time-tracking app that was also called Homebase, which confused some reviewers. On the flip side, my actual experience with Homebase sales and customer service was prompt, friendly, proactive, and helpful. I thought they could do more to further simplify their product and service offering, but as a customer I would easily have considered buying from them.

Some of the limitations on the product and services surprised me, which did shift my distribution to the left, but did so for all competitors not just Homebase. However, it did help me understand further that their solution was really more appropriate for larger buildings, which made me revisit the distribution channels that each competitor was building to make sure that they accounted for this reality. Latch, for example, had gone public partnering with Tishman Speyer, a very large, high-end real estate developer and owner. This was a perfect match for what Latch was trying to do. Homebase had developed a few key distribution partnerships as well; while they might also be good, the uncertainty (to me) was higher than with the Latch / Tishman partnership. To me, this both spread out the mass of the Homebase distribution a bit and shifted it slightly to the left.

Based upon the information provided, however, I now felt I had enough information to move into position sizing. Let’s turn to that now.

Step 04: Position Sizing

The thing about money – at least for the purposes of this discussion – is that it’s a continuous quantity. Therefore, I think a much better question than whether or not to invest is how much to invest. If the amount you invest is zero, that’s fine too. But zero is a quantity just like any other.

Now in many situations it may not make sense to invest $10. Perhaps there’s a minimum investment. You have limited time and attention so you may decide to limit yourself to a fixed number of investments you can track. My point is only that, in general terms, the amount you invest should be a reflection on the confidence you have in the bet, how much it’s worth if you’re right, and how much money you have to be betting with in the first place.

Investing is both a quantitative and qualitative process. Doing the initial assessment of the unit economics, market and company growth rates, and relative valuation were quantitative. Assessing the general market dynamics, consumer psychology, competitive landscape, risk and opportunities were most qualitative. When thinking about position sizing, we move back to the quantitative side. What I like to do here is based upon the information I’ve collected so far, come up with a ‘gut level’ estimate of the right amount of money to invest. In this case let’s call that $X. I then like to use quantitiative methods to calculate an answer and see how far apart they are. If they’re pretty close, I feel good. If they’re way apart, it means I need to reevaluate my assumptions (or my math). I had come up with an initial $X, so now it was time to dive into the numbers.

To assess the amount to invest, I use a modified version of the Kelly Formula. I’ll write a separate post on that at some point, but for those of you who aren’t familiar with it, given the appropriate inputs, it tells you what percentage of your portfolio you should invest in a given bet if you want to maximize your geometric rate of return over time. While a full explanation is beyond the scope of this post, what’s important for now is that it requires you to make estimates of two key parameters:

  1. The probability of success
  2. If you win, how much you’d get for every dollar you invested

Let’s start with the probability of success. Using the odds form of Bayes’ Rule and making a judgement based upon everything I had learned up until now, I made the following estimate:

So I assume a 25% chance of success. Not great odds, but for a startup that’s not bad.

Next I needed to estimate how much I would receive if the company was successful. Using some data on the distribution of returns for similarly staged companies (which I’m sorry I don’t have the link for any more) and defining ‘success’ as any company that returned greater than 5 times my money), I took the expected value of the remaining part of the distribution. This left me with roughly 14x invested capital. In other words, assuming I only looked at companies that had made their investors at least 5 times their money, if I had invested in all of those, I would have made about 14x the money I invested.

Plugging that into to my modified version of Kelly’s Formula:

According to this formula, I should take 6.5% of whatever money I have to invest and invest it in Homebase. Now there are a few caveats here. The first is the ‘whatever money I have to invest’ part. I typically think of this instead as “whatever money I’ve allocated to invest in risky, early stage companies. As you can see from my general investment portfolio, this is actually relatively small.

[The second point is more technical for those who are already familiar with Kelly: though many practitioners use half-Kelly as a general rule, in a 1997 paper Thorpe himself described how to adjust the Kelly approach when multiple opportunities are offered. Beyond this diversification adjustment, there is also the degree of confidence you have in your assessment of the probabilities and payoff involved. You also have a time value of money when investing for an extended period. Finally, when investing in an illiquid asset you lose liquidity and hence optionality. Given this, in my view a 50-66% reduction in the amount allocated seems reasonable. More to come on this in a future post.]

Step 05: Adjust as New Information Comes In

One should always be open to changing their mind as new information comes in. The beauty of Bayes’ Rule is that it formalizes this approach.

In the previous step I had estimated the probability of success (which I defined as a return of 5x my money or greater) to be 25%. However, recall from the very beginning of this post that the shape of the right side of the probability distribution was influenced by the chance that Homebase might be able to go public via SPAC. Since first starting to investigate the company, however, the SPAC market had since cooled off due to some comments from the SEC. In my mind, this definitely reduced the probability of a ‘quick, big win’ scenario. In my estimation, I thought that given the company failed, the probability that this had happened was about 25% higher than if the company had ended up being successful. Therefore, I adjusted in the following way (note that the posterior probability from round 1 becomes the prior in round 2 – this is how Bayesian updating works):

As can be seen, the net effect of this was that my estimate of the probability of success dropped by 4%.

The other thing I did to collect additional information was to ask other people I respect their opinions and concerns. While many folks were very bullish, a few folks were skeptical and provided their reasoning. In general most of the differences were due to higher or lower weighting on different factors, but at least one person did provide me with a perspective that I hadn’t considered before. Net, I adjusted as follows:

This further dropped the probability of success to 17%. Let’s stick that number back into my modified Kelly formula:

The recommended allocation has dropped to 3.7%. This may seem small, but if applied to each position in a portfolio, that means the entire portfolio is only 27 positions; many people would still call that a fairly concentrated portfolio.

How did this compare to my initial “gut level” investment amount? It was lower by a reasonable-but-not-crazy margin, which told me it was probably just protecting me from some of my risk-seeking tendencies.

Step 06: Position Shaping

This step isn’t always possible (or appropriate) in the way I’m going to talk about here, but I wanted to include it because it illustrates how I think about investing more broadly.

Your investments are really just an expression of your beliefs about the world. As I mentioned earlier, your position sizing is part of this expression. In many cases then, I find it helpful to first express my general belief about the world in words and then figure out how to express that in terms of positions.

While a full synthesis of my beliefs at this point would make this post even longer, for my purposes here, there were a few key points:

  1. I fundamentally believed in the continued growth of the smart apartment market
  2. I also believed that all these players – Latch, SmartRent, and Homebase – had tapped into a model with fundamentally good characteristics, strong unit economics, high switching costs, and land-grab/oligopoly dynamics, some network effects, and recurring, SaaS-type revenues.
  3. Each had slightly different strategies, however, on how they thought best to create and capture value. None of them were obviously wrong to me and – despite my efforts – I couldn’t get a lot of conviction about which strategy was the best.
  4. From a valuation standpoint Homebase was the most attractively valued but I could also understand why. And, unlike the public companies, an investment there was illiquid.

Given this, I did what made the most sense to me: I bought all three. This helped mitigate the risk that I would pick the wrong one, but allowed me to make a bet on the sector. Homebase was still the largest position for several reasons (valuation, deal terms, and acquisition potential), but in aggregate I collectively bet on the space with a weighting towards a particular investment.

Conclusion

So there you have it. While long, it hopefully gave you some idea about how I approach investments. Questions and constructive criticism welcome.

A Better Bayes’ Rule

(Let me say upfront that this post has a little bit of math in it. For those of you who are not mathematically inclined, stay with this: it’s is a very useful trick that you can do in 10 seconds and will help you make better decisions. I’ll explain everything and keep it simple.)

Let’s say I’m considering investing in an early-stage startup (Company X) and I want to assess the probability that it will succeed. One the one hand I know that most early startups fail, so investing in them is always risky. On the other hand, this particular company company seems to have a lot going for it, so the evidence is compelling. How should I weigh these two things?

Answer: Bayes Rule. As a refresher, Bayes’ Rule allows you to answer two related questions: (a) what is the probability of ‘x’ being true given some evidence; and (b) if I had a prior belief about the probability of ‘x’ being true, how should I update that believe given new evidence. It should be pretty obvious why/how this could be helpful.

When I first learned about Bayes’ Rule in college, it intuitively struck me as both extremely important and useful. Over the years I’ve revisited it occasionally in an attempt to really drill it into my brain and hopefully get to a point where I would just naturally use it. But it never quite happened as the mental math involved was just a bit complex for me (I am not good at mental math).

Then one day a couple of years ago I came across the odds form of Bayes’ Rule. And it simplified everything. A lot. I won’t go through a detailed explanation on how it works or how it’s derived (if you want that, see here) but let me just show how I practically can use it now and how easy it is.

Play the Odds

First, a quick refresher on odds for those who need it. Let’s say I think there’s a 10% change that my favorite team is going to win the game (and therefore a 90% chance that they wont). The odds are 10:90 or 1:9. In other words, odds are just p(x will happen)/p(x won’t happen). To convert from odds of a:b back to a percentage, just calculate a/(a+b). In this example, it’s 1/(1+9) = 1/10 = 10%.

OK, with that done, let’s move on to Bayes’ Rule. Saying with the startup example, the odds form of Bayes’ Rule says:

An Example

Let’s go through the example of trying to determine the probability that the startup will be a success:

Start with the ‘prior odds’ or the ‘base rate’: p(success)/p(failure) (i.e. the rightmost term in the above equation). From previous reading say I know that the probability for a seed stage startup being successful is 10%, so the odds are 10% : 90% = 1:9.

Next, assess the likelihood ratio, p(E|S) : p(E|F) (i.e. the middle term above). Let’s say this startup has a strong team, a compelling idea in a large and growing market, seems to have a unique take on the space, and is moving quickly.

p(E|S)
To evaluate the numerator, p(E|S), I ask myself “assume a random startup company ends up being a success. What is the probability that this company had all of the things in place that Company X has (at the same phase in their lifecycles)?” The answer is probably ” almost all”. So let’s say 95%.

p(E|F)
For the denominator of the likelihood ratio, it’s the almost the same question: “assume a random startup company ends up being a failure. What is the probability that such company had all of the things in place that Company X has? (again, at the same place in their lifecycles)” Actually, the answer is still probably “most” – even great teams fail regularly etc. – so let’s say it’s 70%.

So now we have the likelihood ratio p(E|S) / p(E|F) = 95% / 70%, or ~ 1.35 : 1

So now we just multiply:

So the odds that the company will be a success are 1.35:9. To convert that to a percentage, we calculate numerator / (numerator + denominator), so we have 1.35/(10.35) = ~13%.

This means that – given my rough assumptions – I should expect that this company has a 13% chance of succeeding.

Why so low? To get a better understanding you can read the full articles, but I think of it this way: the base rate of success is very low. You have some evidence and you have to evaluate how much more likely that evidence is to show up for successful companies than for unsuccessful ones. In this example we’ve estimated that while almost all successful companies will demonstrate that evidence….so will most unsuccessful ones. For that reason, the evidence doesn’t sway us much away from our base rate.

The good thing about this odds form calculation is that I can do it very quick on a napkin, excel, or (sometimes) even in my head. After I tried it once I found it easy to use. And now I use it all the time.

My Personal Portfolio: 2021

Every once in awhile – I won’t say once a year because that would be more frequent than is true – I like to think about my asset allocation. This is different than thinking about making more or spending less. This is about how I split the money I have across different asset classes. I’ll first share how I thought about this decision and then the resultant allocation. I share my thinking for two reasons. On the one hand, others may find it helpful. On the other, others may be able to point out where I’m making a mistake and help me.

Considerations

Here are some of the factors I considered when I was thinking this through:

Target Outcome

I like using the word ‘outcome’ instead of ‘goal’ or ‘objective’, as I think it forces me to be more concrete. When I use the word ‘outcome’, I’m really asking “what does the (relevant part of the) world look like if I’ve done this well?” In this case, the target outcome is that I can sleep at night knowing that my finances have appropriately balanced risk and return while providing me sufficient optionality for the future. More concretely, this implies a few subprinciples with the following order of priority:

  1. Don’t do anything too stupid to mess up the future.
  2. Ensure a comfortable retirement.
  3. Provide exposure to events that could materially increase wealth.

Optionality

In my view, optionality is provided by cash and highly stable, liquid assets. Public market securities (i.e. stocks and bonds) are fairly liquid. Therefore, these assets ought to get a bit more weight than they might have otherwise. Whereas investing in illiquid assets is a difficult-to-reverse decision, liquid assets allow me to quickly correct myself if I realize I’ve made a mistake or the situation (personal or the macro-environment) changes.

Avoiding the Worst Case Scenario

To help plan, I wanted to define the worse case scenario and then, as a first priority, make sure that the chance of the worse case was near zero.

I first defined the ‘worst case scenario’ as comprising either of two outcomes:

  1. A significant loss of capital such that I wouldn’t be able to go at least 2 years without working with no loss in lifestyle and without dipping into retirement savings.
  2. A material change to my retirement outlook (which I predefine as starting at age 65), such that my public securities portfolio alone (which has significant historical returns data) wouldn’t be able to support my “minimum acceptable” retirement income with at least 95% probability.

#1 implied that I wanted to maintain a fairly large position in cash (or other conservative assets). #2 implied, for at least a portion of my public securities positions, a combination of high diversification and a clear and explicit tail hedge. While these two things were likely to damp returns, they provided downside protection.

Allocations

Reserve Account

One of the first things financial advisors always recommend is having a ‘rainy day’ fund: cash enough to maintain your standard of living for at least six months. I took that view and modified it in several ways.

First, because of the relative rarity of the roles I’m likely to accept, I adjust the six months to twelve. Further, because one of my most likely routes is to start another company (which not only takes time without income, but also initial business expenses), I wanted to plan for a two year timeframe.

Now two years worth of cash for me is quite a bit of cash to be sitting on. Having that much sitting in a bank account (at near 0% yields at the time I’m writing this) didn’t seem to make a lot of sense. Money markets weren’t much better. CDs were an option but the yields there weren’t that good either and had penalties for early withdrawal. What I really wanted was a place where I could park my money that was liquid, could earn a reasonable return, and would be relatively safe.

Now I’m fairly risk-seeking by nature so what seems ‘relatively safe’ to me may seem risky to others. That said, I ended up deciding to assign 80% of the value of my ‘reserve account’ to an allocation close to Ray Dalio’s All-Weather portfolio. This has averaged about 8% returns but more importantly has shown relatively minimized drawdowns across economic cycles. Even during the Great Depression (when the S&P 500 lost almost 65% of its value), back-tested All-Weather was shown to have lost just over 20%. Is 20% meaningful? Absolutely. Is it a risk I’m comfortable with? Yes, particularly because I still hold 20% of my reserve in cash and have a tail hedge position (more on that below).

Public Equities

Next in my allocation is public equities. Here I’ve divided my allocation into three strategies.

First is my allocation to net-nets. If you’ve never heard of them, they are basically deep value stocks where the total value of the company (as implied by the stock price times the number of outstanding shares) is less than an estimate of how much an investor could get if they bought the whole company and liquidated its assets. In essence, buying dollars for pennies. Now the availability of attractive net-nets varies so my allocation to this strategy will vary with time, but 10% represents my target position.

Next is my allocation to a concentrated buy-and-hold portfolio. This follows Warren Buffett’s adage that “it’s far better to buy a wonderful company at a fair price than a fair company at a wonderful price”. One of the main reasons for this is tax-advantages: if you can let a company’s stock price compound for many years without selling, you avoid the loss incurred by having to pay taxes when you sell. (Of course, there are many other reasons as well.) Here my target is a 10% allocation to no more than 7 companies that I expect to hold for 10 years or longer. The high bar is intentional, as it forces me to focus and think through what I’m doing. Of course I may hold for less than 10 if circumstances (or my views) materially change.

Finally is an allocation to a widely diversified portfolio of both US and international stocks. This is primarily a nod to the fact that I don’t know what I don’t know. Here, my focus in on simply buying broadly diversified exposure through low- or no-fee ETFs.

Tail Hedge

My current allocation includes a 3% position in TAIL, a means of providing a tail hedge on my portfolio. My view on this is that it allows me to be more aggressively invested most of the time, since when a big downturn happens (such as in March due to COVID), my tail position will offset it. Moreover, during these big drops after the hedge spikes, I can sell off a portion of it and use it to purchase stocks a bargain prices when appropriate. I consider my 3% position a bit larger than what I’d probably do normally, but I am also generally concerned about valuations broadly in the market and believe the combination of historically unprecedented support from the Fed and likely further shifts in both monetary and fiscal policy create higher-than-usual concerns from me about a rapid, negative shock.

Real Estate

I’m a big fan of residential real estate: as businesses go, it’s relatively simple to understand and operate; it’s fairly localized; residential in particular is relatively recession proof (people still need somewhere to live); there are significant tax advantages; and appropriate leverage is easy to come by. Most importantly, however, my wife and her family are very handy: she has assembled her team – agents, brokers, bankers, accountants, and repair folks – that de-risk our investments substantially. For those willing to do the work, owning real estate can be an attractive investment; even for those that aren’t, consider REITs or a private real estate platform like Roofstock or Fundrise (disclosure: I am both a fund and equity investor in Fundrise).

Angel Investments

As an entrepreneur, I feel fairly comfortable making early-stage startup investments. I approach this asset class with a few things in mind. First, due to the power-law nature of returns in startups, I focus on building a large portfolio of bets: 100-125 at any given time (consider AngelList funds for broad diversification). Second, I try to minimize the amount of capital gains tax I need to pay via Qualified Small Business Stock or through investing via my Roth Solo 401k.

New Business Ideas

Beyond traditional angel investments (which are passive), I also am always continuing to think of new business ideas that I may want to start myself or in conjunction with others. Doing so often takes some amount of money, so I have an allocation to this pool as well. I do this to provide myself with some boundaries and make sure I don’t just spend all my money on my own ideas that I fall in love with. While investments in this bucket obviously take the most time, they also provide the most upside potential.

Cryptocurrency

Finally, there’s cryptocurrency. I know very little about cryptocurrency. That said, some investors I respect are quite bullish on crypto and I do see the incredible potential it (along with blockchain) has; a few years back Fred Wilson at USV recommended the ‘average investor’ have a 3% allocation to crypto; as I tend not to invest in things I don’t feel I understand, I further reduced my allocation to 2%.

Summary Table

Given all that, here’s where I end up:

Asset BucketAllocation
Cash2.5%
Non-Cash Reserve: All-Weather Allocation10.0%
Public Equities: Widely Diversified12.5%
Public Equities: Concentrated Buy-And-Hold10.0%
Public Equities: Net Nets10.0%
Public Equites: Tail Hedge3.0%
Real Estate: Private Residential25.0%
Angel Investments10.0%
My Own New Business Ideas15.0%
Cryptocurrency2.0%
Total100%
Asset Allocation

There are a few things to comment on here:

Allocation to Equities

First, when you include the equity position within the All-Weather part of the portfolio, 42% of this portfolio is made up of public equities. While less than the 60% in a typical 60/40 portfolio, public equities still make up a sizable portion of the portfolio. (Bonds, in contrast, make up a much smaller part. Part of this is that I’m attempting to acheive diversificationt through other asset classes. The other part is that I don’t feel I know as much about bonds, and I try to invest in things I feel I understand.)

Second, though there are still sizable allocations to illiquid investments like non-REIT real estate and angel investments, over 60% of this portfolio is still liquid.

Third, a comment on the public equities positions. In reality, these allocations are a bit more fluid. While I intend to keep the diversified holdings invested at all times, the net nets positions depend on the availability of such deals. These tend to get more rare in hot markets (like we’re in as I write this), so deals are not always to be found. The same holds true for the concentrated positions: since here I’m looking for “great companies at a fair price”, sometimes those can be more difficult to find. In those cases where I’m ‘building up’ reserves to allocate to those other areas, I plan to split that reserve 50/50 between cash and the all weather portfolio.

Finally, Insurance

Something I didn’t talk a lot about here but should be part of every investor’s “portfolio” (and is part of mine) is insurance. As you can tell from the above I’m a big believer in managing downside (and tail) risk, and that’s exactly how I view insurance. Beyond health insurance, I have term life insurance sufficient that, if I die and the proceeds are invested in an all-weather portfolio, the income generated from that would be sufficient to take care of my family. In a few more years, I expect that my net worth will be such that we can self-insure and let the insurance expire.

Conclusion

So that’s how my assets are allocated, along with the reasoning that got me there. It lets me sleep at night, allows me to focus where I feel most comfortable, balances risk and return, and feels appropriately diversified. I’m sure it’s far from perfect but it works for me.

Circle of Competence

“The thing about investing is that the size of the circle is irrelevant. What matters is that you know where the edge is.” Applies to all kinds of decisions.