The case for a Customised Benchmark for retail investors.

 The Case for a Customised Benchmark (b/m) for retail investors.

Sometimes when you start to write something, you have the conclusion already planned out well before finishing. The conclusion is set the rest is filler to get there and explain the reasoning. The following note was not like that, it evolved.

When considering a benchmark from a why perspective, I would say to measure your investing performance. Many would agree that is important, but maybe some don’t. fair enough.

Then we consider what to measure ourselves against, that is where the reasoning becomes more complex. Most of the academic literature focuses on accountability, measurability, investible and relevant etc. which is important but means don’t construct an irrelevant benchmark, or a benchmark you can fiddle with it. Fair enough.

When I think about the millions of investors and their individual choices to sit along the risk/reward spectrum, the thought of just one suitable index for everyone starts to become quite puzzling to me. Institutional investors running against a benchmark chosen by their client(s) are quite different. The underlying client may have many managers and equally many benchmarks. The client decides the risk/return trade-offs using many benchmarks and managers.

For the individual investor, my view is that there are three issues to consider in a benchmark to consider and decide what you want to be measured against.

These are the issue of individual stock sizes in the index, secondly sector size in the index and lastly the overall risk positioning in the index. To me, the first two are quite straightforward forward while the third opens up a whole range of issues.

By stock sizes in the index, I mean the weight given to each stock in the index. As we know the market capitalisation-weighted indices give larger weights to the largest companies. In some instances, the whole index can rest on a few very big companies. The question is do you want large companies to be significant in the index? Should the index be exposed to stock-specific risk? The factors the market-weighted index is backing here are size and maybe success. Does that make sense to every investor? Another way to look at it is why have stock-specific risk in the benchmark and if so what stocks? Unless you specifically want that it doesn’t make much sense to me. The conclusion I came to was to use an equally weighted benchmark where every stock has the same impact on the benchmark, and none has the size to drive the returns on its own. Although this outcome is quite desirable and rational to me, it has further ramifications further into the process.

The second is a bit more complex. Does your benchmark carry an over-exposure to certain sectors? Do you want that overexposure? In Australia, that sector is resources. One upshot of taking this approach is to prune the number of stocks that are in certain sectors to even out with bias. Once we have constructed a benchmark to have even stock-specific and sector-specific exposure we can use this to measure our skills against. One issue still hangs over this process, perhaps the hardest one.

When we think about the risk/return trade-off we see a sloping curve from bottom right to top left. The higher the risk the higher the return is the theory. Of course, the variations around that trend line, as shown cleverly in Howard Marks's work, means it is not just a no-brainer to simply add risk and generate high returns.

The concept here is that everyone can consciously choose the level of risk they want to take. They can construct a benchmark of equally weighted stocks and sectors, and that benchmark should, as much as possible, reflect their position on the risk/reward line. The difference then between your performance and the benchmark is due to skill and luck, not risk or style bias.

The variations of benchmarks are endless and accord with your desire to bear risk. The complexity comes in classifying stocks so that the mix of stocks in total properly reflects your desired risk tolerance. That takes a reasonable level of stock knowledge of what to include and what not to. The flexibility of this process is enormous when done right. An individual's benchmark could be customised to measure against his appropriate investment universe. That may include and exclude companies that are very high risk, could also include international companies for instance. Your benchmark could evolve, at 20 years old your benchmark could and should look quite different to you at 70 years old.

Although it took a while to construct, I now run a customised benchmark. That benchmark has 500-600 equally weighted stocks and includes a large swag of international stocks. The largest issue I had was the poor quality of the ASX200. As I equally weighted the Australian benchmark I realised that reweighting the large safe companies raised the risk of the benchmark above where I was comfortable. Unfortunately, the Australian market is not full of quality industrial growing companies. That led me to use international stocks that have many large quality growth companies. That raised the quality of the benchmark to where I wanted it to be. I specifically added a small speculative component. The basic weights were 60% profitable Australia, 30% Quality international and 10% speculative Australia. I purposely purged many small resource stocks to control the resource weight.

This benchmark aims to measure the relative performance of my portfolio. I don’t know the performance of the ASX200, DOWJON or NASDAQ but have stocks from all those indices in the customised benchmark. There will be a reweighting once a year, but change is expected to be evolutionary, not large.

An implicit result of this reasoning is that comparing the performance of different investors who have quite different risk tolerances and that is reflected in their actual portfolios has limitations. Those limitations include accounting for risk, style bias as well as the market conditions that the measurement is made over, ie risk-off or on. Getting true like-for-like comparisons are quite hard. The proper use of a customised benchmark makes much more sense to me.

When someone asks me what my portfolio did? I will reply it did +/-% against my benchmark. That completely bamboozles them. lol