From a stock-screening perspective, investment trusts are a source of both great temptation and frustration. On the one hand, they offer the tantalising prospect of being able to buy shares at less than net asset value per share (NAV). The proposition is made all the more alluring by the fact that, in most cases, a trust’s NAV is made up of investments in liquid and easily realisable assets, usually equities. This seems very much like getting something for nothing.
However, realising the apparent value on offer is not always easy, especially when using screening strategies. Discounts between share prices and NAVs often widen rather than narrow. And aside from betting on the skills of an individual fund manager and the prospects of the market that he or she invests in, there are few judgements investors can make based on solid fundamentals. From the stock screening perspective, this can be deeply frustrating.
But, based on my analysis of 10 years of data provided by stockbroker Winterflood Securities, I’ve managed to find a simple, fundamentals-based, investment trust selection system that has generated significant outperformance over the past decade. Indeed, the average cumulative total return from this system, which I have called the Overlooked and Outperforming trust screen, for the two 10-year periods over which I have tested it (to 31/12/13 and 1/7/14) is a hearty 279 per cent compared with 132 per cent from the FTSE All-Share and 94 per cent from the MSCI World index. The average annual total return from the 20 12-month periods tested is 16.5 per cent compared with 10 per cent from the All-Share and 9.1 per cent from the MSCI World Index.
A question of value
When it comes to screening investment trusts, a key challenge is that there are so few valuation measures available and so much diversity in the sector. In fact, dividend yield and discount are the only really meaningful basic valuation measures investors have. What’s more, yield is not a good way to draw comparisons between all investment trusts because many funds have a policy of reinvesting dividend income to achieve better NAV growth while others are structured to have very ‘lumpy’ payouts.
The discount in itself is also a far from perfect tool for drawing comparisons between different trusts. Wider discounts are normally associated with unpromising investment themes and managers that have a poor performance record. Fortunately, though, investment trust analysts do have a way to square this particular circle with something called the Z-Score, and the analysts at Winterflood Securities have provided me with their accumulated Z-Score data stretching back to 2003.
An investment trust Z-Score should not be confused with the Z-Score used by equity investors to measure the financial health of companies. In investment trust terms, the Z-Score compares the difference between the discount a trust’s shares currently trade on compared with the historic range of discounts the shares have traded at over a given period.
A negative Z-Score suggests a trust is cheap compared with the historic range of discounts while a positive one suggests it is expensive. Importantly, from a screening perspective this is a standardised measure of how cheap or expensive a trust is, which means comparisons can be drawn from the relative value on offer from trusts with highly variable characteristics and discount histories.
However, like most valuation metrics, a Z-Score is unable to differentiate between a trust that is ‘cheap for a reason’ or just ‘cheap’. This is a common problem for investors and for stock screens.
Given that a stock screen is unable to engage with nuanced arguments about a trust’s prospects (for example, a change of fund manager, improved investment outlook in its target markets, and so on), a tangible indicator of quality is needed to use in conjunction with the Z-Score to make this more than just a useful filter.
A question of quality
Normally when an equity investor wants to look for tangible signs of quality they can go straight to a company’s ‘fundamentals’. For example, some of my personal favourites indicators of ‘quality’ when it comes to stock screening include operating margins, return on equity, forecast earnings revisions and dividend cover. These are not particularly sophisticated measures of quality, but it often pays to keep things relatively simple when conducting a stock screen.
Unfortunately, though, as far as investment trusts go, these kinds of fundamentals either don’t exist or, if they do, have little real meaning in the context of a trust.
True, dividend cover can be calculated for trusts using what is termed ‘income EPS’. But as I’ve already ruled out yield as a very helpful comparative measure of value, this is not really much good to us. What’s more, unless a manager is targeting income-generating assets in his or her portfolio, there is a limited amount dividend cover is going to tell anyone about a trust’s quality.
However, there is one broad measure of quality that is available for any investment with a price history: momentum. It’s probably fair to say that momentum is not considered by many as an indicator of quality in the same way that a tangible ratio, such as return on equity, would be.
Indeed, when you buy momentum you never know quite what the market has latched onto to make the shares perform strongly. So momentum should be regarded as a very broad measure of quality indeed.
For my investment trust stock screening system, though, it is my contention that momentum makes a good enough proxy for other measures of quality. After all, for a company or investment trust to be showing strong momentum, the market needs to believe it is doing something right. And many studies have demonstrated that strategies that focus on buying the shares with the most momentum tend to outperform the market over time.
Rules Of Engagement The world of investment trusts is full of the weird and wonderful; from trust financing legal cases to those investing only in Qatar. What’s more, some trusts are extremely illiquid, meaning investors can face huge bid-offer spreads when buying and selling shares if they can trade in any quantity at all. I’ve therefore designed some rules to try to keep this screen away from very illiquid stock and overexposure to the sector’s real oddities. The rules are: ■ Market capitalisation must be more than £100m. This should ensure reasonable liquidity and it also acts as something of a quality test because the fund in question needs to have been credible enough to have raised a decent amount of money to begin with or have been very good at growing NAV. ■ No tracker or hedge funds. My aim is that this screen beats the market (which over the 20 12-month periods tested it has done handsomely). So any trusts that seek to mimic the market, such as trackers, seem at odds with this overall objective. As for hedge fund trusts, these are too complex and objectives are too variable for them to be a happy fit with this screen, in my opinion. ■ No more than half the portfolio (five out of 10 shares) should be in funds with a niche theme and there should be no more than two niche themed funds of the same type in the portfolio. I’ve put this rule in place to stop performance being heavily skewed to a single investment theme which, while potentially offering big rewards, also means big risks. Trusts I define as niche are those focused on non-mainstream asset classes or sub-sectors such as private equity, debt, technology and biotechnology. Those focused on single countries (excluding the UK) or high-risk economic regions such as emerging markets. I also regard Asian smaller companies trusts as niche but not Asian generalists. ■ No more than half the portfolio (five out of 10 shares) should be mainstream funds at the same time. This rule does not apply to global but it does to other mainstream themes such as trusts predominantly investing in the UK (larger and smaller companies), Europe, US or Asia. ■ All trusts must trade at a discount to NAV. Buying expensive assets is not what this screen is about, and while I do not have any hard figures, in my preliminary testing of this screen I found it was noticeably detrimental to the screen’s performance to include trusts that were rated cheaply on a Z-Score basis but nevertheless traded at a premium, albeit a lower premium than normal. |
Going it alone
Targeting shares in either low Z-Score trusts or high momentum trusts on their own has not been an amazing strategy based on my assessment of the 10 calendar years to 31 December 2013.
My test of these two single-factor strategies is based on applying my ‘Rules of Engagement’ (see box) then constructing portfolios of the 10 lowest Z-Score shares and a momentum portfolio of the 10 shares boasting the best share price performance over the preceding three months.
The shares in both portfolios are held for a year before the screen is rerun and the portfolio is reshuffled based on the results. This is the same holding period for my investment trust screen (Overlooked and Outperforming trust screen) that combines both factors (more of that later).
While neither of the single-factor strategies could be considered show stoppers, they have both produced better average annual returns over the period tested than either the FTSE All-Share Index or the MSCI World index (see table).
Comparing strategies*
| To Dec | 2004 | 2005 | 2006 | 2007 | 2008 | 2009 | 2010 | 2011 | 2012 | 2013 | Av. Ann. Capital Rtn. |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Overlooked | 9.8% | 53.1% | 18.4% | 17.6% | -34.8% | 23.6% | 28.2% | -13.1% | 20.8% | 15.5% | 13.9% |
| Low Z-Scores | 16.1% | 45.7% | 15.1% | 6.4% | -45.1% | 12.1% | 23.3% | -11.7% | 17.4% | 18.1% | 9.7% |
| 3-mth Mom | 20.7% | 33.4% | 6.4% | -8.2% | -40.3% | 12.5% | 33.6% | -10.2% | 14.4% | 10.1% | 7.2% |
| FTSE All Share | 9.2% | 18.1% | 13.2% | 2.0% | -32.8% | 25.0% | 10.9% | -6.7% | 8.2% | 16.7% | 6.4% |
| FTSE All Share/MSCI World | 10.2% | 18.0% | 13.7% | 2.4% | -37.3% | 23.6% | 8.1% | -8.8% | 11.4% | 21.2% | 6.2% |
| MSCI World | 11.1% | 18.0% | 14.2% | 2.7% | -41.9% | 22.3% | 5.2% | -11.0% | 14.5% | 25.7% | 6.1% |
Source: Winterflood Securities/S&P CapitalIQ/Thomson Datastream
*My analysis of the Z-Score, Momentum and Overlooked and Outperforming Trust Screen strategies in the table above and graph below are based on share-price returns only and do not factor in dividends. My assessment of the 20 12-month periods for the Overlooked and Outperforming Trust Screen in the next section of this article is based on total return data (share-price performance and dividends paid out and reinvested).
But while both strategies have outperformed on an annual average capital-return basis, on a cumulative basis (tracking the progress of the investment strategies over the entire 10 year period) the high-momentum trusts have actually slightly underperformed (see graph) the FTSE All Share. That’s because of the disadvantageous timing and scale of periods of underperformance and outperformance.
Ignoring charges and spreads, a low Z-Score strategy delivered a 93 per cent capital return over the 10 years to 31 December 2013, while high-momentum shares produced a 60 per cent gain. That compares with 64 per cent from the FTSE All Share and 47 per cent from the MSCI World index. On an aside, to be fair to momentum as a concept, the strategy of a three-month assessment period followed by a one-year holding period is not a very optimal way to conduct a screen focusing purely on momentum.
The good news is that when the two strategies are brought together they start to produce some really interesting outperformance. Based only on capital returns (my full assessment of the investment trust screen below is based on total returns, which includes dividend payments), picking the 10 stocks based on a combination of three-month momentum and low Z-Score, and annually reshuffling the portfolio has produced a 193 per cent gain (see graph). Meanwhile average annual capital return stands at 13.8 per cent.
Source: Winterflood Securities/S&P CapitalIQ/Thomson Datastream
Better Together
To bring together my valuation criteria (Z-Score) and quality criteria (three-month momentum), I’ve turned to the method used by US hedge fund manager Joel Greenblatt to construct his ‘Magic Formula’. The way this works is to rank all trusts for cheapness based on their Z-Scores and for quality based on three-month momentum. The two rankings are then added together and a final ranking is derived from this combination.
I’ve tested the screen over 20 12-month periods, based on the 10 years to 1 July 2014 and the 10 years to 31 December 2013. The graphs on page 29 chart the cumulative total returns (share price movements and dividends paid and reinvested) that would have been achieved by applying the strategy over the full 10 years. The graphs include both the performance of the strategy ignoring dealing costs and bid-offer spreads, and the performance if a 1.75 per cent annual levy is applied to account for both factors.
Hopefully the 1.75 per cent cost represents a fair assessment of this high-turnover strategy (about 95 per cent of the portfolio is changed in an average reshuffle). Based on research from Winterflood, bid-offer spreads on global trusts are normally below 1 per cent for funds with market capitalisations above £100m, dropping to a few basis points for the most liquid shares. While all trusts must have a market capitalisation of over £100m to qualify for this screen, my screen does contain trusts with more exotic investment mandates than those of the global trusts, so I have gone for a more conservative estimate of the likely average spread of 1 per cent.
The remaining 0.75 per cent of the levy is to take account of stamp duty at 0.5 per cent and broker fees, although the hit from broker fees will vary depending on the type of account and the size of investment. The impact of adding in these annual charges is to reduce the 10-year cumulative total return from the screen to 1 July 2014 from 305 per cent to 247 per cent and in the case of performance to 31 December 2013 the 10-year total return drops from 253 per cent to 201 per cent.
Source: Winterflood Securities/Thomson Datastream
While overall performance looks impressive, as is often the case with stock screens, the consistency of outperformance is not amazing. In fact, the screen underperformed the FTSE All-Share in six of the 20 12-month periods monitored and it underperformed the MSCI World index in five of the 20 periods. That said, there were few periods of really grievous underperformance.
The worst year in absolute terms for the strategy was the 12 months to 31 December 2008 when it produced a negative total return of 33.7 per cent compared with a negative 29.9 per cent from the FTSE All-Share and 40.1 per cent from the MSCI World Index. In relative terms, the annus horriblis was the year to 31 December 2011 when an 18.1 per cent total return from the screen represented a 6.9 per cent underperformance of the blended FTSE All-Share and MSCI World indices. Meanwhile, the best year in both absolute terms was the 55.1 per cent total return in the 12-months to 31 December 2005 and in relative terms it was a 33.9 per cent outperformance of the blended indices in the year to 1 July 2010 when the screen delivered a total return of 46.6 per cent.
| Average ann. Total Return | 10 yrs to 1 Jul 2014 | 10 yrs to 31 Dec 2013 | Average |
|---|---|---|---|
| Overlooked | 17.1% | 16.1% | 16.6% |
| FTSE All Share | 10.0% | 10.1% | 10.0% |
| MSCI World | 8.5% | 9.7% | 9.1% |
| FTSE/MSCI blend | 9.2% | 9.9% | 9.6% |
| Cumulative Total Return | 10 yrs to 1 Jul 2014 | 10 yrs to 31 Dec 2013 | Average |
|---|---|---|---|
| Overlooked | 305% | 253% | 279% |
| Overlooked with costs | 247% | 201% | 224% |
| FTSE All Share | 133% | 132% | 132% |
| MSCI World | 95% | 94% | 94% |
| FTSE/MSCI blend | 114% | 113% | 113% |
Source: Winterflood Securities/Thomson Datastream
Top Ten Trusts
As with all stock screens, the results from this screen can be considered either a kicking off point for investment ideas or can be used as a hard-and-fast investment system. I will rerun this screen annually (at this point in the year) in my weekly stock screening column in the magazine. For now, though, here are the current 10 top-ranking investment trusts. There is a distinctly Asian flavour to the results and biotechnology is also well represented, so this is hardly a low-risk portfolio. Trusts covering single themes have been bunched together in the write-ups below.
