If you’re interested in investing and you haven’t heard of SPIVA, you need to put that right. SPIVA stands for S&P Indices Versus Active, and for me it’s the go-to scorecard for active fund managers. Craig Lazzara, global head of index investment strategy for S&P Dow Jones Indices, was recently in London, and I asked him to summarise everything about SPIVA that TEBI readers ought to know.
What’s the idea behind SPIVA?
SPIVA began in the US in 2002, and we’ve internationalised it since then. We wanted to compare, as the name suggests, index performance versus active. So we accessed a database of mutual fund data and then tried to compare funds of a particular type against an appropriate benchmark index — so, for example, large-cap US funds against the S&P 500, mid-cap funds against the S&P MidCap 400, growth against growth, value against value, and so forth. Then we simply compute every six months and ask, How did the average fund do? How many of them beat the benchmark index? How many of them underperformed? Very, very typically the answer is that the majority (and in some cases the very decisive majority) of active funds underperform. Certainly that’s been the case for the last several years in the US and globally, as we’ve expanded the concept.
Costs have a huge impact on investment returns. Often they’re overlooked, but SPIVA includes them, doesn’t it?
That’s right. The data we’re looking at are net asset values. Those NAVs would include the withdrawal of the manager’s fee and the trading costs that were incurred. Of course an index like the S&P 500 is an unmanaged index — you can’t buy the index directly — so, if you really wanted to be precise, you might subtract six or eight basis points from the S&P’s 500 return. It would make no difference to the results, though, which is why we don’t do it.
SPIVA is also adjusted for survivorship bias, which is not the case with most of the performance data the fund industry puts out. Explain what survivorship bias is and why it’s so significant.
It can be very significant. Let’s say, ten years ago, I have a universe of 1,000 funds. I follow those funds through time. Some of them are not going to do very well. The ones that don’t do very well are likely to go out of business or merge into other funds. So, of those 1,000 funds, ten years ago, maybe today, 800 of them still exist in the same form that they existed in previously. Those 800 are biased. It’s not a randomly chosen 800 — it’s the 800 that have done well enough to survive until today. So the database that we use for SPIVA is survivorship bias adjusted. In other words, if you’re looking at 2005 results, it includes all of the funds that existed in 2005, not just the ones that have survived into 2015.
A big problem investors have in evaluating fund performance is style drift. In the UK, for example, where the FTSE is dominated by a small number of large-cap stocks, most active managers tend to veer towards smaller firms. Because smaller UK stocks have generally outperformed the largest ones in recent years, the performance of those funds looks better than it otherwise would.
Let me back up a moment. Here’s a way to think about the question broadly, of active performance versus passive. The fact is that in 2015, the investment business in the US, in Europe and increasingly in Asia, is an institutional business. The importance of small retail investors is, at this point, more or less de minimis. What that means is, if we look at all of the institutions who are competing, some are not mutual fund managers, so they’re not in our SPIVA data, for example. If you look at all of the institutional investors, it’s more or less fair to say they own all of the stock there is to be owned. They also own all of the bonds there are to be owned. The reason, at the end of the day, why active management is so difficult, and why the SPIVA results are as damning, is that the average of all of the managers own the average of all the stocks. So it’s not possible for one of them to be above average unless another one is below average.
There is no natural source of alpha, or risk-adjusted outperformance. In other words, if I’m an active manager, the only way I can have a positive alpha, is if someone else has a negative alpha. And the weighted average value of my outperformance, and all of the other outperformers’ outperformance, has to be exactly equal to the weighted average underperformance of all the managers who underperform. It all goes from one to the other, and of course there are costs on top of that.
But when managers drift into different market segments, it makes it harder to compare like with like?
Correct. Some managers don’t invest in all available segments. They only invest in large-cap stocks, say, or mid-cap growth or small-cap value. For the managers who own stocks simply within one segment, the argument that I made before is still good — there’s no source of alpha other than someone else’s underperformance. What happens in the case of style drift is this. If you have a manager who generally owns large-cap stocks, it’s appropriate in the US to compare him to the S&P 500. But he may have the flexibility to move down the cap scale and buy mid-cap or small-cap stocks. If they do better that year he will look good as a result, and if enough of the large-cap managers do that in a given year, then the majority may look good.
The problem is, if you were to measure them against the universe that they’re actually choosing from, not against only a big-cap universe, the results would be different — they have to be different mathematically. There’s a great article called The Arithmetic of Active Management, written by William Sharpe in 1991. It’s just four pages long but it’s the best case for indexing you will ever read. Basically what it says is, the average of the managers who own assets has to be equal to the performance of the average of the assets. It’s just simple arithmetic.
In Part 2, I’ll be asking Craig:
- Do active managers in other countries, the UK for example, perform any better than those in the US?
- What does the dispersion of returns tell us about a fund manager’s skill (or the lack of it)? and
- Market cap- or equal-cap weighting — Which is best?