Capital Efficiency (or…another post about leverage)

This is my third post about leverage, you can find the first introduction here and a second, practical implementation, here. In particular, I want to comment this recent paper from JPMorgan that I found really interesting (spoiler alert: I agree with them). I will try to ‘translate’ the paper for the curious reader that has only a basic knowledge of finance; I will then show how poor bastards like me can use the descripted concepts, initially tailored for institutional investors only.

The benefits of investing in a diversified portfolio

Let’s say you have £100 to invest and want to achieve a 7% yearly return. You can buy a Stock Index ETF and most likely reach your goal after a decade; the issue is that your final result will be a combination of years when you gain 14% and years when you lose 10%. And you do not like this volatility. So in order to reduce volatility, you decide to add to your investment a Bond Index ETF, an asset that has low correlation to stocks and lower volatility. The low correlation between the two assets will make your portfolio returns more stable: you will not get an 100% steady return year after year, like interest on a bond, but the swings up and down compared to the average, long term return will be lower.

…Ok. I just realised that if you are reading this blog, there is 99,9% chance you are familiar with these basic concepts. The more uncorrelated assets you add to your portfolio, the less volatile the combined returns. This rule works even better if you rebalance your portfolio regularly…and accept that there is no perfect timing for a rebalancing rule: choose a timeframe that works for you and then stick with it.

Volatility is used as a measure for risk. The best investment out there is not the one that offers the highest absolute return but the one that give you the best risk-adjusted return, the highest return per unit of risk taken. It is easy to say “if I invested in Bitcoin in 2014 I would be rich”; since that year, BTC had multiple 80% drawdowns, are you sure you would stay invested in those periods? High volatility means high risk the investor will not stick to the plan. An obvious way to reduce your portfolio volatility is to buy assets that have low volatility. The issue with this plan is that financial markets are pretty efficient, so low volatility assets usually come with associated low returns. Combining high volatility but low correlated assets allow the investor to build a portfolio with great risk-adjusted returns.

David Swensen, the manager of the Yale endowment fund, found a way to cheat around the low vol/low return rule: buy illiquid assets (i.e. Real Estate, Private Equity funds, Infrastructure Funding). They are called illiquid because they are harder to sell compared to a blue chip stock. If you want to sell Apple shares, you can do it with a couple of clicks on your phone AND at any given time the market is open, there would most likely be a buyer ready to fulfil your order; if you want to sell your apartment, you have to find first a real estate agent, a buyer, a lawyer and 23 other things. Even if the price of illiquid assets theoretically can move each second like a stock, the fact that you cannot check in real time the value of your apartment makes it ‘conceptually’ less volatile. They offer returns comparable to stocks because the investor that holds them receive what is called an illiquidity premium, to compensate for the fact that are harder to sell.

Swensen created a portfolio with better risk-adjusted returns than a simple stock&bond combination (or even stocks + bonds + commodities) but he had to pay a price in the form of illiquidity. If you are running an endowment fund, a fund that is supposed to stay in perpetuity, this is a small price to pay; it makes less sense if you are investing to retire and not to leave assets for your kids and grandkids.

So far we tried to create a portfolio that can generate a target return with the lowest possible volatility. Risk Parity was the first attempt (that I know) to solve this problem in the ‘reverse order’: we want to create a portfolio with a target level of volatility that can generate the highest possible return. More precisely, we want each element of the portfolio to contribute the same amount of risk. In the classic 60/40 portfolio, 90% of the volatility comes from stocks; if you want every asset to contribute the same amount of risk, the overall portfolio risk will be therefore capped by the level of lowest risky asset (in reality is not exactly like this but we can use this assumption for the sake of simplicity). This means that after you choose your asset classes and balanced risks, you will most likely end up with an expected return that is low, probably too low for you.

As AQR puts it, Risk Parity gives you the “best” portfolio but it is practically useless because its return is too low. The traditional 60/40 portfolio takes the risk it needs by departing from this “best” portfolio and instead concentrating the portfolio into risky assets. Instead, Risk Parity solves this issue by using leverage. If the unlevered Risk Parity portfolio has a volatility of 5%, you can bring it to 10% by borrowing additional funds; this will increase less than proportionally your returns because you have to pay interest for the loan.

The Problem

The issue with the “best” portfolio is that is an illusion: it depends on how you measure risk (is it volatility? is it how fat is the left tail of the return distribution?), how passive or active you are in the decision making process (when to rebalance), how you build the correlation matrix and how often you review it. Building the best portfolio is more art than science and in fact no one really mastered it, neither AQR nor Bridgewater. The useful lesson from of Risk Parity is that you learn the benefits of using many more asset classes than the traditional portfolio.

The traditional portfolio worked well in the past because the bond component, while reducing risk, was contributing to the returns. But now bonds, and in particular Govies, yield close to zero: the forward looking picture is dire for this portfolio, because to compensate for the lost bond returns, stocks should carry even more weight…and at current valuations this idea seems quite far-fetched. You cannot reduce the % of stocks because they bring the highest return; you can reduce bonds and invest elsewhere but then you also reduce the best protection you have in case of a market crash.

I see the JPMorgan paper on Capital Efficiency as a way to take the the useful bits out of Risk Parity and apply them to the more traditional portfolio.

What is Capital Efficiency?

There are multiple ways to get exposure to stocks: you can buy single names, you can buy a fund or an ETF, or you can buy a future on a stock index. Futures are a type of derivative contract that provides a buyer with an investment priced based on the expectation of the Index’s future value; it allows traders to buy or sell a contract on a financial index and settle it at a future date. So a future give you full exposure to the Index but what is more important here, investors are only required to pay a fraction of the contract value to take a position (this represents the margin on the futures contract). In this sense, futures are more capital efficient because allow the investor to get the same risk/return profile without having to use all the required capital if they were to buy an ETF. This leverage let the investor use the spare capital to buy another or multiple assets to add to the portfolio.

Think about the mortgage on your house, the typical use of leverage. Here you are exposed to the same risk as the un-leverage situation: that particular house you bought. You have simply amplified the gains…or the losses. If you use leverage to buy another asset, you can alter the risk/return profile and, if done correctly, you can lower your risk while keeping all the upside.

The JPMorgan paper explore ways to get cheap leverage for a specific purpose: increase the return of a portfolio while maintaining a proper risk exposure. While futures are available for retail investors, I strongly suggest their use. First, even in their E-mini size, they still represent a quite big exposure, circa $200k at current price; remember that you want to rebalance frequently, so either $200k is just a small fraction of your stock exposure (good for you!) or you need to build a blend exposure of future + Stock ETF to manage the portfolio. Second, you need to manage daily margin; this means you have to set aside part of your portfolio in cash to be ready to deposit in case of a margin call. If done too conservatively, this means a drag on the overall portfolio return, partially invalidating getting leverage in the first place (I will leave to your imagination what happens if you do not have enough cash to cover a margin call). This strategy makes sense for institutional investors because futures offer very cheap leverage and margins are managed by professionals.

Why Capital Efficiency (or leverage) is important

Let’s see a practical example using two portfolios created by PortfolioCharts.com:

• the classic 60/40 portfolio has an annualised return of 6.3% and a standard deviation (another measure of risk) of 11, which result in a Sharpe Ratio, return per unit of risk, of 0.57
• the All Season Portfolio (a mock-up of Risk Parity) has annualised return of 5.7% and StDev of 7.9, with a Sharpe Ratio of 0.72

The All Season Portfolio is superior on a risk adjusted basis, but since we eat retire on a return basis its use might be un-efficient. If we add a bit of leverage (14.3%) and assume a cost of leverage of 1.5% (what I pay with Interactive Brokers), we can align the All Season levered portfolio to a return of 6.3% and a Sharpe Ratio of 0.69: bingo! We now have the same returns of the 60/40 portfolio with less risk.

What can YOU do

As I said in the intro, retail investor cannot follow exactly what JPMorgan suggest but this does not mean that the paper is useless. First you have to remember that any use of leverage makes your portfolio riskier than the unlevered situation: leverage reduces your margin of safety (the accuracy of your assumptions vs reality) and it is an additional element that has to be managed, either via rebalancing or margin calls. But all the tragic stories that your read revolve around too much leverage, the addiction to it. Archegos, to pick a recent one, was leveraged 20 to 1 while in my example I use 1.14 to 1, quite different. As AQR puts it, proponents of the use of leverage argue that using leverage can be risk-reducing rather than risk-increasing provided four conditions are met: (i) enough unencumbered cash is kept to meet any margin calls (ii) leverage is applied to a well-diversified portfolio (iii) assets can be rebalanced frequently and (iv) counterparty risk is minimized. As we discussed, (ii) and (iii) are part of the central thesis of this post, while I will soon explain why (i) and (iv) are not relevant for a retail investor.

I found and use three instruments to get leverage:

• a margin account on Interactive Brokers. This is my preferred solution because it is easy, you just have to open the account, and cheap, 1.5% per year in UK. You do not have to manage margin calls because you can post your assets as collateral and as long as your collateral is a well diversified portfolio and your leverage is low (I use 1.33 to 1 but I think IB does not allow to go above 1.5) you should not have any issue. Your counterparty in the loan is your broker, so if they go broke…well, you will have other things to worry about.
• a levered ETF. Forget about the 3X ETFs, they are TOO levered to be managed, but I find the 2X version good enough. If your target portfolio exposure to the S&P500 is £100, you can buy £50 of the 2X ETF and use the remaining £50 to buy another asset. The main issue with these ETFs is that the leverage is reset at the end of each day, so to maintain the desired exposure to that particular asset within your limits, the higher the volatility of the asset, the higher the frequency of portfolio rebalance required (and potentially the higher the transaction costs). For these reason, is better for you to pick a levered ETF on a low volatility asset…which is easier said than done because the commercial need for these products is exactly the opposite. This means the S&P500 2X is the ONLY levered ETF I would suggest you to use and only if it does not represent your sole portfolio exposure to stocks (in this case you can lower the frequency of rebalance).
• the NTSX ETF. Called the 90/60 portfolio, this is the Wisdom Tree version of a 60/40 levered portfolio. They created it EXACTLY for our needs, just buy it and send them a thank you note (srsly, they wrote a paper about it, go on an read it before making any decision). Wisdom Tree is also your counterparty exposure risk, this is why I also suggest you to listen to their podcast 😉

Conclusion

I have heard so much whining about low interest rates (professionally and not) that my ears are bleeding. What if we consider this as an opportunity or at least a prompt to do something different? Yes, it is a bummer that to get the same results as twenty years ago we have to work harder…but was it even right to have risk free assets paying 4% when inflation was at 2%? Maybe THAT was the exception and now we are in a more normal environment. Low yields means that we can borrow cheaply, if the situation changes and revert back to what it was, we can always lower the borrowing and shift back the portfolio to the OG 60/40.

I would like to leave the last thought to the p2p environment. The institutional investing world, starting from Swensen, embraced illiquid assets because of their high yield and peculiar risk factors; it is quite telling for me how retail investors, instead of using p2p for the same reasons, are happy to sacrifice part of the returns in exchange for a perceived increase in liquidity. Maybe an entrepreneur will crack the code and provide the Holy Grail of liquid high yield p2p, but I am sure that so far you are better of choosing the certain, high yield and illiquid, for the uncertain, less yield and “*” on liquidity.

What I am reading now:

Follow me on Twitter @nprotasoni