The Fund returned -5.79% during the quarter ended December 2021, 4.48% behind the benchmark MSCI Emerging Markets (Net) Total Return Index, which returned -1.31% for the period. For the year as a whole, the Fund returned -14.75%, 12.21% behind the benchmark return of -2.54%. This has been the worst (relative) performance year for the Fund since inception, and we apologise to investors for this underperformance. While a year like this is both unpleasant and uncomfortable, it is also not totally out of line with the Fund’s history: in 2018, the Fund was 11.66% behind the market (this was followed by 19.76% outperformance of the market in 2019); and in 2015, while the Fund ended the year 6.69% behind the market, there was a point during the year when it was 12% behind on a rolling one-year basis. Calendar year 2015, in turn, was followed by two successive years of 1.0% to 2.5% p.a. outperformance. Over three years, the Fund has now outperformed the market by 2.24% p.a., over five years it is 1.73% p.a. behind and over 10 years the Fund has still outperformed by 0.04% p.a., in spite of the recent tough period. Outperformance since inception stands at 1.22% p.a. over 13.5 years.
The reasons behind the Fund’s sometimes uncomfortable swings in relative performance are multi-fold, with the biggest factors being the high active share (over 80%), off-benchmark exposure typically around 40% since inception, a concentrated portfolio (50-60 stocks), and lastly, the fact that, given our long-term (five year+ time horizon) valuation-driven approach, we are often invested in a number of companies that are disliked or out of favour. Examples of this today would include JD.com (and China internet more broadly), Magnit and AngloGold (all top 10 positions), as well as several others. We also still own 15 of the 20 largest individual stock detractors of 2021 – in other words, over the long term, we believe a large part of 2021’s underperformance will be recovered. The weighted average upside of the Fund today is around 70% (well north of the historical 45%) and the Fund’s weighted five-year internal rate of return is 20% p.a. (also much higher than where it has been historically).
For the quarter under review, the biggest positive contributor to alpha was AngloGold, up 31% for a +68 basis points (bps) contribution to relative performance. Naspers and Prosus contributed +46bps, while NetEase (Chinese gaming) contributed +34bps. The final two material positive contributors were LVMH (global luxury) and Anglo American (diversified mining). These contributed a combined +61bps, split fairly evenly.
As one would expect in a very negative quarter, there were several material detractors. The biggest among these was Sendas, a Brazilian cash-and-carry retailer and overall, the second-largest food retailer in the country, which returned -32% and cost 82bps of relative performance. Aside from general weakness in the Brazilian market, Sendas’ share price reacted poorly to the group's plan to buy 71 underperforming hypermarket stores from CBD, its former holding company, with a view to turning them around. In addition to the high price tag of R$5.2 billion (c.$1 billion), the lack of shareholder vote when the parties are related (French group Casino is the largest shareholder in both Sendas and Assai) was not well received by investors. We lobbied the board to put this to a vote but were unsuccessful, as the group argued the valuation and due diligence were done by independent parties that ruled that the transaction was priced fairly and in the interests of all parties. Despite our unhappiness with certain aspects of this transaction, we understand and agree with the strategic rationale (a significant immediate addition to the store base with generally attractive sites that otherwise would take many years to achieve), and we have retained the position in Sendas. In our view, it remains a very attractive asset (very well run, with a resultant return on invested capital of over 25% and in a category that is gaining market share) on a very attractive valuation (around 12 times 2023 earnings).
Two other Brazilian stocks made up the next biggest detractors. PagSeguro (card acquisition and digital banking) halved in the period and cost 74bps, while XP Inc. (securities broking and wealth management) fell 29% and cost 48bps. Rounding up the top five detractors were Yandex (Russian search, ride hailing and general tech), which cost 44bps and Trip.com (Chinese online travel agency), which cost 37bps.
Ordinarily we would spend much of this piece discussing what went right and wrong in the quarter; however, it is more useful to rather look at the year as a whole to understand why the Fund underperformed by such a significant margin. While in summary it was simply a year of poor stock selection, one can break down the drivers of underperformance in 2021 into five categories that all played a role.
- China education: The single biggest impact on relative performance came from the Chinese tutoring/education stocks. New Oriental Education (EDU) was the Fund’s main exposure here and it returned -90% after the government effectively converted the industry into 'not for profit'. This massive regulatory change was unprecedented and more far-reaching than we had anticipated. EDU cost the Fund 2.87% of relative performance in 2021. A smaller position in Youdao had a -52bps impact, which was largely offset by not owning TAL and GSX (+40bps impact). As we always try to do when we make mistakes that are material, we have taken away lessons from this painful experience.
- China internet: Tencent Music Entertainment (TME) was the second-largest detractor for the year (-1.5% alpha impact) and JD.com the sixth-largest detractor (-69bps impact), while Autohome and Trip.com cost the Fund a combined -84bps. Given this, it is tempting to conclude that Chinese internet stocks had a similarly large impact on the Fund than that of the education stocks. However, the reality is a bit more nuanced, as there were some stocks that we didn’t own at all, were significantly underweight in or sold completely as a risk management measure when regulatory intervention was accelerating in the middle part of the year. As an example, the Fund was underweight Alibaba throughout the year, and this contributed +56bps to performance. Not owning Tencent boosted performance by 81bps, offset fully by the combined Prosus/Naspers long position, which cost 95bps to give a small net negative impact. We also sold Meituan Dianping to zero, which stock had a small positive (7bps) impact on relative performance. Conversely, NetEase (a 2.5% position on average during the year and only 0.5% in the index) was actually up 8% for the year and was a top 10 positive contributor (+43bps). When one adds up all the Chinese internet stocks owned or in the benchmark, this sector as a whole had a fairly small contribution of around -0.5% to underperformance in 2021. We believe that selected stocks in the internet sector in China remain very attractive, particularly when some are either big longer-term beneficiaries of regulatory changes or are not significantly affected by them. For example, JD.com is a net winner from regulation as it was previously negatively affected by anticompetitive behaviour from the market leader, Alibaba. Some stocks, like TME, have seen their share prices fall by far more than their business is likely to be impacted by regulatory tightening, and so we have maintained exposure to them, albeit at lower sizes than before. The Fund has around 23% exposure, either directly or indirectly (via a fair share of Naspers and Prosus), to China internet, reflecting our conviction on the potential opportunity in several names in this sector.
- Country weights: Although the emerging markets asset class as a whole (as represented by the benchmark) returned -2.5% in 2021, this disguises a large range of returns between the various constituent countries/territories. Despite being underweight China, which was the worst-performing major constituent, the impact of China education, internet stocks and Melco Resorts & Entertainment (Macau gaming) cost the Fund 4% of relative performance. The three strongest-performing territories were Saudi Arabia, India and Taiwan. The Fund has no exposure to Saudi Arabia (typically lower quality assets and on high valuations), and this cost close to 1% due to the 38% market return there. Taiwan and India were both up c.26%, each of which cost the Fund 2% of alpha, in large part due to the percentage of the Fund invested in these countries being far lower than their respective weights in the index. The Fund is built from a bottom-up perspective, rather than with targeted country weights in mind, and we did not find sufficiently attractive opportunities in these markets. In Taiwan, Taiwan Semiconductor Manufacturing Company returned 12% in 2021 (for the American depository receipt) and ended up costing the Fund 60bps of relative performance – although we owned it throughout the year, our 4% average position size was lower than the 6.5% average weighting in the index. Although one of the other Taiwanese holdings did extremely well (Momo.com), it was a small position and sold to zero when it reached fair value. India was similarly negative for the Fund: much of the Indian market is expensive, in our view, with most higher-than-average businesses trading at multiples ranging from 40 times to 80 times forward earnings, and without the level of earnings growth that would justify this type of valuation level for us. To compound matters, the stocks we did own in India did not do as well as the broader Indian market. For example, HDFC Bank, which is owned and is not in the benchmark, declined 10% in 2021, whereas ICICI Bank, which wasn’t owned and is in the benchmark, was up 36%.
- Stocks not owned: The Fund, on average, holds 50-60 stocks through the cycle (including a number of which are not index constituents), compared to a benchmark of over 1 550 stocks. This concentration versus the dilute benchmark is a deliberate part of our process of selecting the best investment opportunities on a risk-adjusted basis within our investment universe. Generally, the stocks we don’t own tend to be a positive contributor to relative performance. However, 2021 proved to be an exception to this historical trend. Of the approximately 1 500 stocks not owned in the benchmark, the net contribution to alpha was -3.3%. In five of the last seven years (inclusive of 2021), the impact of having zero weight in the majority of benchmark stocks has been positive for the Fund, which is what one would expect from a strategy that is fairly concentrated relative to a dilute benchmark, and one that has a general preference for the better businesses in emerging markets. The -3.3% is therefore an abnormally high headwind to performance.
- Low cyclical sector exposure: Our investment philosophy focuses on finding undervalued assets, using a long-term investment horizon. Within this, we have a strong preference for good businesses (which tend to be the less cyclical assets), although valuation always prevails. For this reason, we typically have less exposure than the market (through the cycle) to more cyclical industries such as energy, basic materials, industrials and banks (particularly State-owned ones). That doesn’t mean we will never own lower-quality businesses – we will, and do (for example, our new buys during the quarter included an oil company, a car manufacturer and a low-cost airline, none of which we would describe, at face value, as being good businesses), but cyclical assets in aggregate have not been a material part of the Fund. A sector analysis shows that the more cyclical industries hurt Fund performance in 2021. The Fund had almost nothing in energy, costing 1% alpha. The materials, industrials and banking sectors generally had high single-digit/low double-digit returns and these collectively cost the Fund another 1% of alpha.
There were several new buys in the quarter, with the largest new buy being Petrobras (1.7% of Fund at year-end), the Brazilian oil and gas group that we last owned in 2011. There have been a number of corporate governance improvements after the scandals that plagued the firm in 2015 at the tail-end of the PT (Worker’s Party) administration. Importantly, the huge capital expenditure required to exploit the oil finds off the eastern coast of Brazil (the ‘pre-salt’ fields) has largely been spent, making Petrobras highly cash generative at current oil prices. Importantly, this has resulted in a significant reduction in indebtedness, with net debt to earnings before interest, taxes, depreciation and amortisation having come down from a level of four times to only one times over the last five years. The lack of oil exploration/new finds provides some support to current oil prices, and at these levels, the valuation of Petrobras is very attractive, and is still attractive at our estimate of the long-term (lower) normalised oil price. Petrobras trades on four times free cash flow and offers an extremely attractive 20%+ dividend yield. There is talk of further privatisation of the asset, which will further boost potential return (due to less government influence). At the same time, government interference (as has happened historically) is always a risk, particularly with the very possible return of Lula da Silva (of the PT) as President in elections later this year, following the lifting of his corruption convictions. A four times free cash flow multiple and 20%+ dividend yield mean a lot of risk is arguably priced in, but the risk still results in a limit on the ultimate position size.
The second new buy of note (1% position) was Porsche, the holding company with its stake in Volkswagen (VW) being its main asset. VW’s emerging market exposure is over 50%, with China being the biggest part of this through both direct sales and joint ventures (accounted for as associates). The Dieselgate scandal of 2015 is largely behind VW and the firm has made great strides in developing its electric vehicle range, which is arguably among the most impressive of the legacy car manufacturers. Of great attraction to us is the high premium share of profits, with the Audi and Porsche brands making up half of operating profit. Premium brands have more pricing power, generate higher margins (e.g. the Porsche brand enjoys operating margins of 18%) and a higher return on capital than mainstream brands, greatly improving the quality of VW’s earnings. Due to the superior qualities and metrics, the multiple that is assigned to a premium brand should arguably be far higher than that of a mainstream brand. Ferrari (albeit a better business than Porsche and Audi, in part due to its very exclusive and powerful brand, and in part due to controlled supply, not dissimilar to Hermès in luxury clothing/accessories) provides some support to this view, with its rating of 45 times 2022 earnings. Yet VW, with a large part of its earnings coming from premium brands, trades on a lowly six times 2022 earnings, with an attractive 4.5% dividend yield, and Porsche, in turn, trades at a discount to this, given the discount to the value of its stake in VW.
The Fund also purchased a 1% position in Taiwan-based MediaTek, a well-diversified fabless (design chips but outsource production) semiconductor company, with revenue mostly derived from mobile (c.60% of revenue), connecting chips, smart home and power management-integrated circuits. Revenue at MediaTek has grown 19% p.a. cumulatively over the last 10 years, in the process making it the fourth largest in the world and largest mobile chip system vendor by volume, overtaking Qualcomm. MediaTek has navigated the evolution of demand for its products very well, with management having steered the company through multiple product transitions – from optical disk drivers to TVs, feature phones, smartphones, and others. This is in spite of the technical barriers to managing such transitions. The large profit pool of each new area that they have entered has been exploited thanks to heavy research and development on their part. Earnings are up four times over the last two years, driven by 41% annual revenue growth and operating margins more than doubling, from 9% to 21%. We expect mid-teens revenue growth going forward, in spite of the higher base, driven by multiple factors, not least of which is the much higher average selling price of 5G products (relative to existing 4G) and new flagship chips being launched. MediaTek trades on 15 times forward earnings and generates returns on equity of 25%. With free cash flow conversion of over 100%, it also offers an attractive 6% dividend yield.
Other new buys worth noting, but all smaller so far (60-90bps), are Southeast Asian gaming and e-commerce operator SEA Ltd, Chinese sportswear retailer Anta, Eastern European low-cost airline Wizz Air and Russian lender TCS Group. In the case of all four of these stocks, we already covered them but didn’t own them, largely due to valuation, and it was sharp price declines that brought them into buying range: SEA from a $365 share price peak to the current $195 price, Anta from a HK$190 peak to HK$107, Wizz Air from £55 to £40 and TCS from $120 to $79.
To fund the above purchases, a variety of positions were sold. Most notably, we took the decision to divest entirely from tobacco, which was around 3.5% of the Fund in total in four holdings at the beginning of the quarter. We do not believe that our investment universe has been materially impacted by this decision – since inception of the Fund 13.5 years ago, there has been a mean exposure of 2.5% to tobacco stocks and they have had a combined immaterial relative contribution to performance of -0.2% since inception. Three other small positions were also sold entirely, namely the insurer Prudential PLC (consolidating the pan-Asian insurance exposure in AIA), Turkish hard discount retailer BIM (Turkey uninvestable for now, in our view, given unconventional monetary policy) and the Chinese online vehicle portal Autohome (to add more to higher-conviction China internet stocks without increasing overall China internet exposure).
Please note that the commentary is for the US dollar retail class of the Fund.
Please note that this is for the retail version of the fund. View the Global Emerging Markets Fund page