Using Leverage to Reach Your Financial Goal (Sooner)

Ok, the title of this post feels a bit thrashy but stay with me for a minute. In the end, it might be really that simple.

I spent a couple of weeks working on a video about leverage with Giovanni (the final result is here, unfortunately it is in Italian). My recent post about the increase in the cost of using leverage picked up Giovanni’s attention; after recording a video on 0DTE options, we thought we could revert back to easier topics…therefore we devoted half an hour talking about ergodicity. Guess what: even if you are the only one on the entire Italian YouTube to use an abstruse word in the title, it does not bring you that many clicks. Shocker.

While the concept of leverage was firmly stuck in my head, I read this from Nick Maggiulli. Another crazy idea popped into my mind: what if I borrow his style and describe what a moderate use of leverage can bring to an otherwise pretty standard portfolio? Simple and yet novel. I never read something similar and I tend to pay attention to this kind of content.

Pretty soon I realised why he has a gazillion followers (and a cool job too). It is hard to do what he does. Not impossible but not a one-hour task either. With a new job, two kids and an international move that is not yet finalized, I do not have much time to spare (maybe if I stop watching every NBA playoff game…). ChatGPT offers great drafts that help to turn complex topics into more digestible bites of information. What I really miss is the data: gone are the days when I could download a price series from Bloomberg with a few clicks. With nothing to throw into Excel, I reverted to another familiar tool: Portfolio Visualizer.

We live in a complex world, each one of us has our own circumstances. The key to writing the perfect piece is to nail the correct assumptions and reduce the universe of possible scenarios to a manageable number of indicative cases. In Nick’s post, each decision leads to a specific result because everything is deterministic: the hypothetical wanna-be pensioner relies only on their savings to retire. The math behind each scenario is easy. Once we introduce the option to allow savings to generate a return, via a highly volatile asset like stocks, all those tables go out of the window. Unless you consider periods of 30 years or longer, when it is (almost) safe to assume that those returns converge to their historical, arithmetical average.

Considering the complexity of managing leverage, employing it on a portfolio that does not consist predominantly of volatile assets is a bit useless; instead of leveraging a bond-like portfolio, just buy a stock index: you would get a similar risk/return profile without much of the headaches. [Unless you found a great, sort of relative-value/arbitrage strategy, in that case write me an email 😉] So, to be clear: if your portfolio is less than 50% stocks, stop reading, buy more stocks and come back in a year.

How to we build meaningful scenarios?

Using only the past arithmetic average of returns is stupid. Otherwise, this post would have just been a line: “US stocks returned 10% / year on average so as long as your cost of leverage is lower than that, go for it”. If you believe in the equity risk premium, the correct sentence would be: “use leverage as long as the equity risk premium is higher than the spread above the risk-free rate you pay on your borrowings”

The maximum drawdown and the volatility of the investment strategy are as relevant as its returns. We care about the destination AND the path to get there. Using only the past as guidance is stupid because the past represents only one of the possible paths that could have emerged out of the same distribution.

We need a Monte Carlo simulation. If we have a portfolio of assets, we need assumptions on their correlations and in which specific scenarios their correlation increases. We need assumptions on how past returns and drawdowns are likely to be indicative of the future. Think about survivorship bias or the 20s crash for US stocks: will the US alone continue to outperform, or even perform in line with a wider basket of stocks? Will we experience again an 80% drawdown?

I do not have a Monte Carlo simulation, so I have to try to squeeze something not-as-stupid out of something stupid. But hey, anything in here is investment advice, innit?

Help!

There are a couple of considerations that can help here.

The whole idea is to start to use leverage early. And by early, I mean when an investor has enough understanding of what they are dealing with. Samuel of Picture Perfect Portfolios recently described how he acquired great knowledge about investing in just one year, starting from zero. Sure, he had to dedicate to it more than four hours a day, not a trivial amount of time; but the key message is that we are not dealing with rocket science, a bit of willingness to learn will get there anyone in a reasonable timeframe.  

Using leverage means that the portfolio % swings will be bigger than the unlevered scenario. Since we are early in the journey, it means that those swings will involve smaller absolute amounts. Crucially, the levered investor will experience something that the unlevered investor would also eventually face, it will just happen earlier. And if they manage to stick with it, it will happen more frequently.

As I wrote in the introduction, I want to analyse moderate leverage, up to 1.5x maximum. Meaning that a 24% drawdown will turn into a 36% drop (that’s what happened in 2022 to an unlevered vs 1.5x levered SPY portfolio according to Portfolio Visualizer): the difference is not trivial but is not that rare either for SPY to lose more than 30%, it happened 6 times in the last 60 years.

Facing this loss, the levered investor has two choices:

• They are spooked by the drop and want to sell. Great, they learned what’s their REAL risk appetite and know they cannot hold even an unlevered 100% stock portfolio going forward. Better to realise this sooner rather than later and re-adjust their saving rate or their goals.
• They stick with the strategy and march on, to the next, bigger drawdown 😉

The 1.5x levered investor needs a 66% drawdown to be wiped out. This loss is unlikely but not impossible for a well-diversified, large-cap stock portfolio. But there is another aspect that comes to the investor rescue: they are regularly adding to the portfolio, they are in the accumulation phase. And again, they are early in the process.

The levered investor is dollar cost averaging and by doing so they are regularly adjusting both their portfolio purchase price and margin. Investment losses are lower, in % terms, because they are constantly adding to their portfolio at lower and lower prices. And in absolute terms, their Net Worth is still mainly driven by their savings and not by their invested assets (here Nick Maggiuli again, explaining this concept better than me).

The leveraged investor is simply accelerating the moment when Investments will dominate Savings. AND THAT’S WHEN THIS POST STOPS! I am not suggesting running leverage indefinitely, that involves a different conversation. I just want to show that using moderate leverage to reach the point when Investments and Savings pull the same weight is not extremely risky.

Playin with Portfolio Visualizer

Portfolio Visualizer is a great tool but has some limitations:

• there are only US-listed instruments
• you cannot back-test indexes

While the first one can be circumvented by European investors (if you are interested in testing an ETF on S&P500, use $SPY in PV and then invest in $SXR8), the second one is more annoying because ETFs are fairly “recent” instruments: for example, in the last 30 years we had only a year (and half) of high inflation, 2022 (and counting). It is difficult to test a broad number of scenarios, the necessary type of exercise to confirm the robustness of an investment strategy.

PortfolioCharts is not the perfect tool either because the guy behind it sells his Excel models but not the datasets (obviously…). It is still useful to check historical volatilities and drawdowns for certain unlevered portfolios: if, for example, you are interested in running a levered 60/40, it is useful to know that (according to HIS dataset), the worst drawdown happened in 1974 and was not that different from the loss in 2022 (the returns in PC are net of inflation!).

There are two ways to simulate leverage in PV.

The first one is to run a portfolio that has positive weights >100% and then square it back to 100% assigning the corresponding negative weight to an ETF like $BIL. $BIL is an ETF that invests in a 3-month Treasury Bill, so its return matches in an (almost) accurate way the cost of leverage embedded in futures.

Here is an example of a 1.5x levered 60/40 portfolio, as represented by $VBAIX:

The result is encouraging but…this represents the life of a futures investor (or the cost embedded in a levered ETF like $NTSX). If I plan to use leverage via a margin account, like the one offered by Interactive Brokers, I have to increase the costs. Unfortunately, I cannot simply slap a margin on top of $BIL (well, if you know how to do it, feel free to contact me :)) but PV allows me to input the Debt Interest. The average Fed Fund rate in the considered period was 0.92% (see here) and if you add the 1.50% margin charged by IB you arrive at 2.42%. The result is not 100% fair to reality because the Max Drawdown would take into account this average cost of leverage and not what was paid up to that point but it will not dramatically change the picture.

Every metric is worse compared to the previous test but the portfolio is still performing decently. You can visually observe the effect of leverage in the “angle” at which the portfolio grows, or declines. After almost 10 years, the levered portfolio starts to increase faster, meaning the effect of compounding is gearing up. Even considering the 2022 decline, the distance between the two lines is currently bigger than in 2020; if you remember PortfolioCharts’ data, 2022 has been the biggest shock since the 70s for the 60/40: even that type of shock didn’t derail the levered portfolio.

Let’s stretch the analysis using the oldest available ETF, $SPY. Considering that I am playing with a riskier instrument compared to the 60/40, I reduce leverage to 30%. Again, the average Fed Fund rate in the period has been 2.42% (see here), which brings us to 3.92% including the IB margin:

The Sharpe and Sortino ratios are (surprisingly?) quite stable and even the Max Drawdown does not differ massively between the 2 strategies. I have added the Log Chart because it makes more evident the dynamics between the levered and unlevered strategies; if you just look at the non-log chart, the levered portfolio might get undeserved points. Few considerations:

• The $SPY ETF was listed at the beginning of a massive run for US stocks; starting a leveraged position back then was, in hindsight, not bad timing
• the levered portfolio went below the unlevered one during the 2008-09 crash

If we think the assumption that “stocks go up in the long run” is true, which should be a given otherwise there would be no reason to discuss whether to use or not leverage, the conclusion to the above points is that moderate leverage works:

• any starting point after 1993 does not change the Max Drawdown
• the 2008-09 crash was, on a real basis, on par with the worst drawdown in the last 50 years (see PortfolioCharts). This crash did not…crack the strategy
• after 30 years, the levered portfolio is 300k (or 23%) bigger than the unlevered one

It is not time to declare victory yet.

A proper analysis would run many 20-year period simulations (why 20? With 10 the difference between the 2 strategies would be small anyway, with 30 the levered S&P500 will always win unless we up considerably the volatility, i.e. we introduce crashes like the 1920s…but again, if that’s your expectation, I guess you do not touch stocks even in the unlevered fashion anyway) and look at a risk threshold like the worst 95th percentile. As said, I cannot show you that.

This is the worst I can find with PV:

The low is in Feb-09 when the levered portfolio is worth 197k and the benchmark 226k (-13% on a relative basis). It takes a bit more than a year for the levered portfolio to reach and surpass again the unlevered one.

And so?

With this post I wanted to provide some food for thought, a starting point for the reader to experiment and think about their preferences, risk appetite and circumstances. There are many other risks to consider when dealing with leverage: for example, I do not think that any PV analysis will show you Black Monday in 1987 because the crash happened in a single day. If that’s going to happen again, I bet IB will knock on your door for some additional collateral.

If you are (still) interested, I would strongly suggest you check my other posts about leverage.

What I am reading now:

Follow me on Twitter @nprotasoni