This is the latest in a series of commentaries by the European family office firm.
Here is another in a regular series of commentaries by Blu Family Office.
We always make it very clear to people when we talk about our investment process, that we take no macroeconomic views or, more generally, we do not try to predict the future development of asset prices. Rightfully so, people often ask, well if we don’t make predictions then what do we actually do? Quite a bit, as it turns out. It all has to do with discerning fact from fiction and using sound statistics to manage our risk.
But let’s start at the beginning: every investment model needs input. As we all know, “excrement in, excrement out” and so too, we have to make sure that our data is clean. For example, historical prices, global population growth or gross domestic product are examples of clean input. Earnings forecasts, analyst’s ratings, or anecdotal evidence are not. Why? Because the former are facts that are indisputable and in the public domain, whereas the latter are someone’s opinion and subject to change. You cannot build a rational model with input that is not reliable, just like you cannot build a house on a foundation that may or not be altered entirely.
We make sure that the data we have has a verifiable track record and that we have access to data that comes out in the future. When we put together our model for our global equity index, for instance, we started with more than 100 sets of data streams but could only use 18 in the end, given the disparity and inaccuracy of historical numbers. Golden rule here: when in doubt, don’t use the data. It is much better to have 18 perfect inputs, than 100 somewhat iffy ones, when trying to engineer statistics with a very small margin of error.
Even more important is it to understand what one is looking for in the available data. We live in the post data revolution world. Data is everywhere, in fact there is so much data that we now have fake data and fake news. Clearly, everyone has caught on to using data mining, quantitative modelling and advanced game theory algorithms that use data to great effect.
So, naturally, the next evolutionary stage, is now the dis-engineering of what was a brief utopia of free and reliable data available everywhere when the internet age first started.
Ironically, this has made being human more important than ever and a critical sanity check in making sure we apply another layer of scrutiny, as we discern good data from bad. An easy example is the high correlation between ice-cream sales and drownings. True, they go hand-in-hand with each other, as people who drown may have also eaten ice-cream on a nice sunny day. But it doesn’t mean you drown from eating a cold fluffy mixture of egg, milk, sugar, and flavouring. Moreover, we eliminate news in general and any other contrived analysis, meaning someone else has done something with the data. Keep it simple. Remember, all we want is clean data, and that should be pretty easy to find once you eliminate all the obvious bad input.
Now that we have made sure that everything coming in is clean, we can begin to do things with the data. This is the most crucial point in the investment process where immense creativity, persistence and diligence is required so that value can be articulated. It is where one has to scrutinize whatever theory one applies in trying to use the data to one’s advantage. Thanks to the wonders of using machines, we can also test whatever model we come up with, versus events that happened in the past.
Most people don’t give much credence to so-called “back-tests” as reliable indicators of a strategy’s success when traded for real. While we whole-heartedly agree, they do give us a wonderful tool to better understand the behaviour of a chosen investment strategy. This is the trial and error stage of our investment process, and we want to get this right (and wrong) before we devise a plan and define a strategy to implement.
For us, this means using advanced mathematics and quantitative methodology to look for patterns, trends and confirmation of behaviour to categorise risk and enable an efficient allocation of capital. Ultimately, it is the output that we believe is the natural conclusion from the sound analysis of proven and reliable data.
Once we have defined our strategy and we know what we want comes the implementation. We need to get exposure to those investments that can deliver the best return for the risk we take. This means scrutinising the fees and making sure we are not paying for any unrewarded risk in our allocations. It is akin to playing the role of a doorman at the hottest club or restaurant in town. We discern any and all investment strategies as to their underlying risks and drivers of returns, e.g. why do they make money? We check both qualitatively and quantitatively to confirm that the strategy delivers the risk exposure as promised. It is a surprisingly easy yet brutal selection process, only letting in the well-heeled guests through our doors and making sure any imposters are left by the wayside.
Finally, we define clear risk parameters for taking profits or taking losses and understanding how big a bet we want to take when allocating our capital to each investment. We may not be able to foretell what the future will hold, but we can most certainly decide how much risk we want to take. Most important is being consistent and unemotional about the investments we make. The rest is just numbers and getting rewarded for the chances we take.