Managers: can they predict?
IPE’s Investment Manager’s Expectation Indicator displays predictions on different asset classes of approximately 122 asset managers and is published in every issue of IPE. Russell Investment Group, in the data pages each month, summarises the total figures of how many managers are positive, neutral or negative. This summary gives an insightful overview of how the financial community looks at the most import asset classes.
But is there more to learn from the individual opinions? This issue will be discussed in a series of articles, starting with this article on the individual qualities of different asset managers for ‘equities’.
In the last few years, analysts have received quite a bit of criticism, primarily because the equity markets fell through the ceiling in the years 2000, 2001 and 2002, while they kept most equities on Buy, especially in the beginning of the bear market.
In the Financial Analysts Journal (2003) two papers were published where analysts showed significant herding behaviour for internationally diversified conglomerates, and that they had a penchant for smaller growth stocks, while in the recent bear market they actually should have given more attention to the ‘dull’ value caps. On the other hand, it should be noted that during the bull market years, analysts received much less criticism, as research showed the stock chosen outperformed in bull years.
However, far less research has been done with regard to the overall asset class visions of the analysts/strategists of the major investment houses. Perhaps this can be attributed to the fact that data providers have not maintained quality databases of the predictions of strategists. IPE has been tracking investment manager’s expectations on a monthly basis since February 1997. Taking into account the approximately 120 asset managers with varying lengths of history, we come up with more than 7,000 monthly predictions, which means we have a terrific database for assessing the value of the manager’s expectations.
In this first article we dive into the equity expectations for the five countries/region’s IPE keeps track of: US, Euro-zone, Japan, Asia and UK. We split the database into two parts. The first parts consists of managers that have contributed more than seven years of views, which, depending on the asset class, numbers about 55-60 managers. This means they have contributed their views in both the bull and bear markets.
The second part consists of all the managers that contributed around two years of opinions. In the beginning of 2004, IPE added a considerable number of managers to their Investment Managers’ Expectation Indicator. These managers are not included in the first part of our research as there might be an element of luck involved when only 20 views or less have been expressed. These views can, however, be used in a broader context when we wish to find out whether or not managers in general add value.
In IPE’s Investment Manager’s Expectation Indicator, managers offer their predictions for the coming six to 12 months. In our methodology every monthly prediction is considered as separate for both 6 and 12 months; so we calculate the percentage rise (or fall) of the index for both the 6 and the 12 month period, and this on a monthly basis according to the following formula:
Dj,k,t * Rk,t
j = manager
k = the asset class/region
t = time period
Rk,t = price return (over the 6 or 12 month period)
Dj,k,t = -1, 0,_1 with Dj,k,t = 0 if the manager expects no price change (neutral), Dj,k,t = -1 if the manager expects the price to drop and Dj,k,t = +1 if the manager expects the price to increase.
The overall manager score is then calculated as:
Dj,k,t * Rk,t
The monthly six and 12 months scores were than averaged. Next, we ranked the n managers from 1 for the best manager to n for the worst manager on the basis of these average results for all five equity regions defined by IPE individually. The top-performers overall for the asset category equities were the ones with the best average rankings (equally-weighted), based on our five regional analyses.
Via regression models, we tried to explain these return prediction result ranks using the following factors:
q Rank in assets under management (size);
q Rank in % of equities in total of assets under management (EQ%);
q Rank in % of equities in US, Europe, Japan, Asia and Emerging Markets respectively (as percentage of firm assets under management: US%, EUR%, JAP%, ASIA% and EM%);
q Rank in return of the equity strategy views for the other equity country/region and total views (RETUS, RETEU, RETJAP, RETASIA, RETUK, RETTOT) so as to test if there are links between prediction results in the various regions.
We also incorporated two dummy variables, one for the country (country) and one for the style (style) .
The transformation into ranking variables helped us solve problems related to non-normal distribution of some variables. This was especially a problem for the skewed assets under management variables. It did not apply to the dummy variables for country (US = 1 and Others = 0) and style (Quantative/Structured = 1, fundamental/micro = 0, mix = 0.5).
We extracted the information on these variables from IPE’s July/August supplement issue of the Top 400 Asset Managers and their corporate website. The equities indexes we used come from MSCI. Both univariate and multivariate regressions were used to detect patterns in the score and to distinguish between good, mediocre and bad managers.
In general, we can conclude that the 59 managers do add value, albeit quite modest (0.29% over period of more than seven years). This figure was considerably higher when adding the managers we omitted earlier (those with approximately two years of history): 1.83%. Most markets were in a steady, upward trend in 2004 and 2005, and most managers have been (correctly) optimistic in the last two years.
Certain managers were able to add a lot of value, while others performed quite badly; we also find that the top 10 per region differ substantially. There seemed to be a dichotomy between the US, UK and Europe on the one hand and Japan/Asia and emerging markets on the other. The ten best managers (on average, with at least seven year track record) are shown in table 1.
Results of univariate regressions
In table 2 the results are shown for univariate regressions. The first thing we can conclude from this is that SIZE is not really an issue to determine the score within a univariate framework, perhaps with the exception of Japan where we see something of a ‘boutique effect’. Another thing that is apparent in a univariate setting is that the %EQ (the percentage of equities of the total assets under management) is a negative factor, except for the UK. This means that relatively more equities under management leads to poorer results, so specialising on equities is not really beneficial, even when deriving views on this asset class itself. Apparently, the valuation of the average equity market has a level of complexity that demands a broad multivariate analysis that also takes into account other asset markets within and outside the country.
The univariate regression also proved that the COUNTRY effect was not significant, which may contradict the adage that Americans tend to be better asset managers. This could be attributed to a relatively small number of American houses. From our research we cannot conclude that Americans are better asset managers; nor can we say that Europeans are worse. It is known that many American specialist boutiques have good managers and strategists. These are not included in the list, whereas the best of the smaller European houses are in the list.
Another thing we can conclude from the univariate regressions is that there is indeed a high correlation between the predictive powers for the separate regions. Those who are good in the US tend to be good in the UK and Europe, and vice versa. The exceptions are Japan and Asia, who seem to need a different, specialist approach.
The last conclusion we can draw is that style matters; structure-oriented managers that have a quantitative approach to their investment processes are better, especially in UK and Asia. Fundamentalists can compete with the quants in the US.
We have come up with some nice preliminary conclusions from our univariate regressions, but we had hoped to explain more.
Do multivariate regressions where the combined effects of the variables are tested change the picture? All above-mentioned variables were taken into account, together with the percentages of how high the percentages of equity under management in different regions are. We have also looked at assets under management value . The following themes could be distinguished:
q Correlations between market visions. In the multivariate framework the intra-equity allocation variables turn out to be an important, direct link with the US, UK and Europe. If you understand the UK and Europe, the probability of being a good US manager goes up significantly (and vice versa). In the univariate regression it proved that there was no significant link between market visions between Japan/Asia and the US (the only non-significant link in this respect!).
Running a multivariate regression, we see that things are actually a bit worse in that it hurts (with regards to predicting the US) to be ‘Big in Japan’ or an Asian specialist (ie, the market driving forces in Japan/Asia on the one hand and the US on the other differ to such an extent that too high a concentration/ focus on the valuation model of the one market hurts your predictive power in the other). The main driving forces differ by sector; bottom up seems to be more important in the US while elsewhere in the world, country factors and intermarket relations are more important. But things get more complicated when analysing the links between the east and the west. Whereas in the US it didn’t help in any way to be good in Japan or Asia when selecting good American strategists, the opposite is not true: Unlike what we saw in the univariate framework for Japan, predictive power in the US is a positive factor in a multivariate setting.
Good Japan managers tend to be Japan specialists (see SIZE factor in the table) with the better ones being those that read the American market well. This corroborates our earlier finding of differences in the driving forces behind the market mechanisms in the US and Japan. (International) macro factors and top-down analyses are of more importance when trying to unravel the secrets of stock market trends in Japan than they are in the more bottom-up/sector oriented US market. Just like the US, Europe and the UK seem to be a related sub-sample when it comes to market visions and their quality, and the same applies to the subset Japan/Asia;
q Relevance of the country factor. In the US and Japan the country factor (being from American origin or not) did not play an important role. In a multivariate model, the big American houses do not seem to do well in Europe; this is not to say that there are not smaller houses (boutiques) with good track records in Europe, however. Although the country factor was not significant in a univariate regression, we see a remarkable shift when we move to a multivariate framework. With the US managers in our sample (on average) being relatively large compared to the average European (or other) as managers in the sample, the country factor shows up significantly, thereby suppressing the overall rank of the large US houses. The opposite is true for Asia, where American asset managers do a much better job than Europeans;
q Relevance of the style factor. In a multivariate framework this factor still appears to be significant for some regions, but for the US we can say that there are more roads to Wall Street;
q Size. When studying size patterns in our sample, the relative diffusion of results is striking (see table 3). In the US this is not a very important theme. In Europe, however, in a multivariate framework, size matters with larger firms in terms of AUM doing a superior job. However, it might seem logical, after our univariate regression that this would translate into the percentage of equities being larger as well, as it now seems to be the case that this univariate factor (percentage of equities) was a mere proxy for the percentage of European equities. This means that the more diversified, large houses do a better job in predicting trends in European equities. Size matters in Europe, but not if this size stems from being big in your own equity market.
In Japan, however, firm size seems to work against the manager. Japanese equities under management tend to be a positive explanatory variable. Thus, one should not go for the big asset manager factory, but for the big Japan specialist instead. In the rest of Asia size is not an important factor per se, but what is important is that good managers tend to concentrate on more than one asset class (EQ% negatively correlated just like in the univariate regression). When looking at size in the UK, one aspect of this factor clearly dominated the others; size hurts!
In conclusion, we can say that the managers included in the IPE’s Managers’ Expectation Survey can indeed add value, and on studying the results more carefully we come up with some very interesting results which certainly merit more research for the other asset categories in the survey, such as bonds and currencies. For equities we found that size was not that much of an issue (even two negatives for Japan and UK). Size seems to matter in Europe, but not if there is a high percentage of European equities under management. Apparently you should also understand the other asset classes and probably also the flows between them. It should be noted that our multivariate framework performs less good (ie, less explanatory power) in Japan/Asia than in other countries (see table 4); these markets do seem to have their own peculiarities.
The overall explanatory power seems strong enough to warrant incorporation of these results into some kind of manager’s selection context, though we can, of course, not translate strategic visions directly into a strategy performance of asset management products. However, with consultants and clients all being aware of the fact that performance in the past is not necessarily a flawless indicator of future success, information concerning the quality of the manager’s strategic views will always be useful information when selecting managers. It will help distinguish between those managers that are fortunate to have achieved good performance and those that truly deserve their excellent track record.
We do therefore conclude that the IPE’s Managers’ Expectation Survey is useful, as long as we keep track of the link between manager signals and their investment implications in a structured manner.
Erik van Dijk is chief investment officer and Harry Geels is senior investment manager at Compendeon, based in Bunnik
in the Netherlands
The authors wish to thank Bruno Solnik for his helpful comments