Bernstein Passive Investing Serfdom Aug 2016

23 August 2016 23 August 2016

Fund Management Strategy

The Silent Road to Serfdom: Why Passive Investing is Worse Than Marxism Inigo Fraser-Jenkins +44-207-170-5134 [email protected]

Paul Gait +44-207-170-0599 [email protected]

Alla Harmsworth +44-207-170-5130 [email protected]

Mark Diver

Policy makers should care about active fund management because of the role it plays in allocating capital in the economy. This action is a force for social good and indeed comprises the social function of active management (and for that matter of sell side equity research). This note aims to help asset managers and asset owners make this social case for active management to policymakers. Passive management has been a hugely beneficial development in lowering the costs of investing. To be clear up front we are in no way anti-passive. Simple factor strategies will soon be regarded as passive too which will lower costs still further and more importantly make it clearer where asset managers are delivering idiosyncratic returns which are truly valuable. This is a good thing for investors.

+44-207-170-5132 [email protected]

However, policymakers have regarded the rise of passive as entirely benign. This is myopic as there are costs to the system overall if active management suffers a catastrophic demise.

Sarah McCarthy, CFA

Ultimately this comes to the social function of active investment. Its primary role in this respect is capital allocation and as such it is a force for social good. A Marxist economy where investment is centrally planned is a plausible alternative but less efficient. However, a supposedly capitalist economy where the only investment is passive is worse than either a centrally planned economy or an economy with active market led capital management.

+44-207-170-5131 [email protected]

Robertas Stancikas, CFA +44-20-7170-0595 [email protected]

Alix Guerrini +44-207-170-5133 [email protected]

Jonathan Absolon +44-207-170-5101 [email protected]

Marion de Floris +44-207-170-0541 [email protected]

Maureen Hughes +44-207-170-0511 [email protected]

We examine in detail the implications of this for capital intensive industries such as mining and demonstrate the crucial role that active management plays in information discovery and highlight the difference that the two approaches to investment make in the real economy. We don’t need to make absurd ad infinitum forecasts of passive management reaching 100% of asset to outline its impact on the functioning of the market. We extrapolate the impact that passive can have on correlation and liquidity long before that point. ESG forms a crucial part of any defence of active. If an asset owner or government sets an ESG target they are, implicitly at least, espousing a belief in the power of active capital allocation as a force for good. Also, in the rush to passive some active decisions are being made but no longer explicitly recognised as such. We raise the question of who should have fiduciary responsibility for factor allocation? As passive assets become so large does this create a natural opportunity for active? This is wishful thinking. (1) It is not clear how one would know that such a point had been reached. (2) We outline why passive AUM will continue to grow (3) active and passive scale differently. Passive needs to be as large as possible while active has real capacity limits. Thus we think there is no point at which one reaches "active nirvana". We suggest that the balance of active vs passive is unlikely to naturally mean-revert and that when policy is formed at the very least it should be ensured that it does not encourage a further shrinkage in active management in a way that would be inimical to society at large.

Analyst Page

Bernstein Events

Industry Page

See Disclosure Appendix of this report for important disclosures and analyst certifications

www.bernsteinresearch.com

23 August 2016

DETAILS We have been writing a lot about the ways in which active investing can be defended against the rise of passive and semipassive (smart beta) strategies1. That research is mainly directed at the management of asset management companies. However, we think there is a bigger picture here, so this note is directed at policymakers. OK so politicians and regulators will probably not read a sell side research note, so we hope that it will provoke asset managers and asset owners to influence governments and regulators. Another way of phrasing this is that our previous note was seeking to defend active management by setting out the case for what individual asset managers needed to demonstrate and what individual asset owners should seek out in terms of characteristics. In this note we defend the role of active management for the system overall, ie at the level of the economy. Active investment decisions form a crucial part of the capital allocation process in an economy and as such there is a clear and distinct social worth in their aggregate action. A possible alternative is a Marxist economy where the capital allocation is planned, such a system is perfectly viable but just less effective. However, a supposedly capitalist economy with no active investment – where passive management is the only capital allocation process – is, in our opinion, worse than either of these alternatives, hence our assertion in the title of this note. The commonality between both active market management and the Marxist approach is that in both cases there are a set of agents trying – at least in principle – to optimise the flows of capital in the real economy. It is just such a feature that is lacking in passive investment management. Of course it is a matter of debate as to whether the set of individual participants that compose the active market are collectively better at allocating capital than the technocrats that dominate command-control economies. But given that Western economies have entrusted capital allocation to the market the implicit abrogation of that responsibility through the rise of passive management, without the establishment of an alternative mechanism for capital allocation, is an insidious problem. There are some bigger issues at stake than the miniutiæ of how to differentiate between a given active and passive fund. Up until now policymakers have regarded the rise of passive as a benign force. In fact, to the extent that policymakers have thought about the issue at all (and most will not have done) we sense that they have looked favourably on it as a way of lowering the upfront cost of running national pension systems. We think this is short-sighted and one of the principal points of this note is to suggest to policymakers that they also have an interest in the role that active management plays and that the rise of passive is not an unmitigated benefit. This comes down to the question of the impact on the financial system of the rise of passive investment both in terms of its potential impact on financial stability and also its ability to fulfil the important role of capital markets for society at large. This is not a simple issue and we will address several different aspects. 

The role of active in helping the allocation of capital in the economy



The importance of capital allocation for society



The need to have rising asset markets to fund the pension system



The role of active as a source of liquidity and how passive magnifies correlation shocks



ESG as an implicit expression of belief in active management



What happens to fiduciary duty in a world where the active-passive distinction is no more? Who gets the fiduciary responsibility for factor allocation?



Is there a limit to the size of passive or a natural equilibrium between active and passive?



Does this need policymakers to sit up and care or can the industry self-correct?

Let's be clear up front. We are not saying that passive is a bad thing. Far from it. The growth of passive has delivered a major benefit to asset owners in lowering the cost of access to equity market exposure. When this cost cut is applied to simple factor strategies as well, which is now happening, and so-called smart beta becomes available at passive rates this allows a potential further cost saving. We think that the benefit to asset owners goes even further than this. If the market and all standard factors (Value, Growth, low vol, Quality etc) are available at passive rates then the buyer of a fund can more easily determine what they 1

Please see In Defence of Active Management

FUND MANAGEMENT STRATEGY

BERNSTEIN

2

23 August 2016

are paying for in terms of systematic versus. idiosyncratic exposure. Put another way, it allows asset owners to be clearer about exactly what it is that they want to pay for. This was the subject of our previous note.2 However, there are a few broader points to consider. As an asset owner one needs to be able to know how to benefit from this cost reduction, if it is only achieved by the asset owner taking on the risk of making explicit asset and factor decisions themselves (ie in deciding what smart beta index to buy) then it is a false economy. Such asset owners could always have just gone out and bought single stocks but presumably they wanted to hire an asset manager as they want their skill in equity investment. The question of factor allocation is the Achilles heel of smart beta and the question of who should have the fiduciary responsibility for factor allocation has not been addressed in enough detail. The main focus of this note, however, is a broader question. If we go to a more macro level than the individual asset owner-asset manager relationship what are the implications of the switch from active to passive for the system overall? Some would suggest that because the average net-of-fee return from active management is less than that for passive that the fee paid for active management is a net drain on society. This is a non sequitur. A given investment in active may or may not be the best decision for an individual particular investor but for the system overall there is a benefit in the efficient allocation of capital. This can come directly through the provision of equity capital or in forcing the dissemination of information that is crucial for the raising of capital through lending or credit markets. That is what we ultimately seek to show in this research note. Ultimately this goes to the heart of the question, what is the social function of active management in equity markets, and indeed of sell-side equity research? In the wake of the financial crisis we think it is even more important than normal to demonstrate that there is indeed a social function. A field of endeavour that performs no social function is ultimately unsustainable if it has a cost that is imposed on the rest of society. Any such activity will, in the ultimate analysis, simply be regulated out of existence. However, there is a clear and distinct task that active management (and, by extension, sell side research) performs. This is in the allocation of capital either directly through the raising of capital in primary markets or else indirectly in the information discovery process. This is a laudable task and needs to be recognised.

THE PROLIFERATION OF INDICES AND WHAT DO WE MEAN BY ACTIVE VS PASSIVE? Before we start our discussion proper we should say what we mean by active and passive. It is becoming increasingly hard to tell the difference between active and passive. More than a decade ago things were easier as passive meant market cap weighted indices, although even then the number of indices was growing. The proliferation of indices has been greater since then as smart beta indices increasingly offer exposures that cover part of what was known as active before. We now have the bizarre situation that there are more indices than there are large cap stocks, this is not at all helpful for investors, Exhibit 1.

2

Ibid

FUND MANAGEMENT STRATEGY

BERNSTEIN

3

23 August 2016

EXHIBIT 1: The proliferation of indices 6000

5000

4000

3000

2000

1000

0 1920

1929

1960

1985

2010

2012

2016

The first 5 datapoints are based on Wurgler (2011) “On the Economics Consequences of Index-Linked Investing” The last two datapoints refers to the cumulative number of factor indices (4274 from ERI Scientific Beta (http://www.scientificbeta.com/#/concept/homeanalytics-intro) and 673 ETF’s identified by Morningstar (http://www.ft.com/cms/s/0/a5309ec0-43dd-11e4-8abd-00144feabdc0.html#axzz4Ek414pmW) We have fitted an exponential curve though we have left the scale on the x axis non linear on purpose as in fact the recent rate of index creation exceeds that fitted by an exponential curve. Source: Wurgler(2011), FT, ERI Scientific Beta

In a sense we are in a period which has seen a triumph of benchmark indices through the combination of passive asset flows and that remaining active funds tend to be constantly assessed relative to a benchmark. However, in another sense there are now so many indices that the original idea of a simple benchmark seems lost. How many indices does one need? We think that already there are more indices, particularly smart beta ones, than can be of use to investors, but the rate of index creation does not appear to be abating. This is in part because it is unclear what the recipe is for commercial success for smart beta indices, so index providers, ETF platforms and others are competing to just create indices and see what works. A consequence of this is that we think that it is actually no longer possible to point to an absolute boundary between active and passive, the distinction is now more subjective and in the eye of the beholder in that it depends on why a given strategy or index is being bought. Yes, we accept that the cap-weighted index is in a sense the only true passive index as it is the only index that all investors can buy, but declining costs of smart beta mean that it will soon be possible to buy smart beta for the same fee as traditional passive. Also if a strategy follows a simple pre-defined and transparent rule with no discretion to, say, buy cheap stocks then is it really any different from an index provider in a traditional way forming an index from the 100 stocks that are largest on market cap in the UK for example and calling that the benchmark FTSE100 index? In fact such an index could be regarded as more passive as far as transparency goes than the Dow, for example, which has discretion in its construction. One could possibly distinguish between the strategy formation and the implementation and judge a given approach as active or passive with regard to the latter. There also used to be a cost difference but that is rapidly being eroded and is set to go away altogether. We think that the most simple smart beta strategies will soon be available for the same fee as a traditional market cap weighted passive index. In that case there is no ability to distinguish between active and passive in a simple way on cost grounds either. The answer, we think, is that we don’t think that investors should get too hung up on whether a given strategy is passive or active in the traditional sense, there are other distinctions that matter more. For the purposes of this note we assume passive to mean traditional cap weighting and also indices that follow simple, transparent, rules-based strategies. It is the economic impact of the rapid growth of popularity in such indices that we assess here.

FUND MANAGEMENT STRATEGY

BERNSTEIN

4

23 August 2016

ROLE OF ACTIVE MANAGEMENT IN HELPING THE ALLOCATION OF CAPITAL IN THE ECONOMY This note is not about why an active or passive allocation may be optimal for a given investor, instead it is about the role of active management for the system overall and what may happen as more capital is switched to passive management. This comes down to the role that active investment decisions have in capital allocation in the economy which on the one hand is the key defence of finance having a social role and also why governments might need to care about the state of active management. The role of markets in capital allocation is a long-standing concept and by no means an artefact of equity market activity in recent decades. Walter Bagehot in Lombard Street: A description of the money market3 spends most of the introduction to the book making the case that the ability of capital markets to rapidly reallocate capital into expanding and shrinking industries was a key element in the superior economic growth of the UK compared to other countries in the late nineteenth century. What of the role of equity markets in society overall? In 1939 J.D. Bernal published an influential book: The Social Function of Science4. After what was seen as the use and misuse of science during World War I and the Great Depression it was a call for an explicit recognition of the impact that science had on society. The purpose of the book was then to frame the policy question of how science should be organised to result in the maximum benefit to society. Bernal was far more interventionist in his suggested response than would be acceptable or indeed even recognisable today, but post the financial crisis a similar underlying question has been posed of the financial system. What is the social function of equity markets? More specifically what is the social function of fund management and equity research? A key part of the answer has to be, we suggest, the promotion of more efficient capital allocation with a view to it fostering higher growth rates in the overall economy. The causal path is either direct for companies that need to raise equity capital, or indirectly in allowing a dissemination of information that is critical to companies raising capital from banks or from credit markets. Some may reject this as an interesting line of inquiry for it brings us ultimately to moral questions that are not normally the subject of equity research notes. But we think that to ignore this question would be a misreading of the times. In the wake of the financial crisis and with a far greater focus on inequality it is beholden on market participants to make this case. The good news is that we think that such a case can indeed be made. But this is not just a case of having to prove one's social benefit, there is a question for policymakers as to whether they need to take interest in the role of active management for the good of the system overall. We suggest that one could rank order possible societies by their efficacy of capital allocation. We suggest that such an ordering might look like 1)

Capitalist society with functioning capital markets

2)

Marxism

3)

Capitalist society with predominantly passive capital markets

In a Marxist society at least someone is doing the planning of capital allocation, but in a predominantly passive market then the capital allocation process is done by a marginal participant. What do we mean by "predominant"? We have to recognise that no one, to our knowledge has yet been able to derive a theoretical level at which point capital allocation breaks down. Indeed how would we even know when that level had been reached? However in this note we do provide a model of the economy that shows, at least in principle, how the degree of market failure in the real economy can be shown to scale with the degree of nonactive management of the financial economy. Here we take mining as an example of a paradigmatically capital intensive activity where the role of efficient price discovery is the most important (after all, capital light activities by definition are ones where capital allocation efficiency will matter the least) and develop a model of the coupling between the real and financial economies that highlights the principle in question. The 2015 Venice Biennale included a work of art called Das Kapital Oratorio that consisted of actors reading the whole of Marx's magnum opus over the course of several months in a piece of performance art directed by the artist Isaac Julien. It was not, it is true, a very efficient way of digesting the work (your author got to see only Part 2 Chapter 17: The Circulation of Surplus Value) but it was a highly original way of bringing a new interpretation to the classic work. To follow the link in Exhibit 2 below please Ctrl + Click on the picture . 3

Bagehot (1873) Lombard Street: A description of the money market, available at https://books.google.co.uk/books?id=MGYuAAAAYAAJ&dq=editions:v9KYeuq_PbgC&pg=PR3&redir_esc=y&hl=en#v=o nepage&q&f=true 4 Bernal, J (1939) The Social Function of Science, Routledge. This work also happened to have a Marxist undercurrent.

FUND MANAGEMENT STRATEGY

BERNSTEIN

5

23 August 2016

EXHIBIT 2: Das Kapital Oratorio [Ctrl + Click on picture to follow link]

Source: Author's own video: http://players.brightcove.net/2197926683001/r1VSbAudB_default/index.html?videoId=5080843941001

If our angle here is the social function of investment and the framework for capital allocation, how do various alternatives stack up? Marxism sees the process of perpetual capital accumulation under capitalism as its fundamental weakness. As capital gets accumulated to the point where more profit is earned than there are profitable investment opportunities in the economy, this leads to overproduction and over-accumulation of capital. A Marxist economy - or a socialist system which Marx envisaged would come to replace capitalism- would operate according to a very different economic dynamic. Very simplistically, a Marxist economic system would not produce and invest for profit, instead basing those decisions on the criteria of directly satisfying human needs and producing 'use-values'. Marxist 'production for use' would be coordinated through a process of rational planning of use-values and coordinated investment decisions to attain economic goals based on utility not profit. The distribution of output at this post capitalism stage of socialism would be based on the principle of 'to each according to his contribution'. This is in contrast to capitalism where market forces would compel capitalists to produce use values as by-product of the pursuit of profit. The rational planning of production and investment would mean that the overaccumulation of capital - where investment grows faster than profitable investment opportunities and where sectors of the economy get created that do not produce any economic value but are needed to absorb investment, leading to waste, inefficiency, crises, bubbles, recessions and the accompanying social problems – will no longer be present. Although Marx provided little actual detail on how socialism might be organized, he did envisage planning to in some way involve the input and decisions of the individuals involved at localized levels of production and consumption, and Marxist planning does not necessitate the kind of centralized planning of the Marxist-Leninist variety which formed the basis of the Soviet economy. Indeed, there have been numerous other conceptions of economic planning, including decentralized planning or participatory planning. This type of arrangement would have planning and other decision making distributed among various economic agents or localized within production units, rather than having economic information aggregated by a central authority that then plans production and capital allocation as in a command or centralized planning system. Such arrangements are often envisaged to imply some form of democratic decision making within the economy or within firms. Alternatively, some economists have proposed computer- based forms of decentralized coordination between enterprises. Some commentators have argued that this type of system - a non-market type of 'democratic socialism' that is an alternative to both 'market socialism' (a system which socializes the means of production but retains market competition) and Soviet type central planning - allows for a self-

FUND MANAGEMENT STRATEGY

BERNSTEIN

6

23 August 2016

regulating system of stock control to come about, based solely on calculation in kind, and refutes the objection that any large scale economy must rely on a market system of prices in order to function. Of course, even a high-level discussion of the merits of the various forms of socialism that have been proposed, both relative to each other and relative to capitalism, is way beyond the scope of this report. It seems uncontroversial enough to say that the Soviet Union has been vastly less efficient at allocating capital than that of the USA and other western capitalist countries (see for example Exhibit 9). However, there is no a priori reason to assume that an economy based on Marxist principles implemented through a very different system of economic planning cannot be a viable economic model. Further, we would posit that it is plausible that a rational, forward-looking planning of investment would be superior to no planning and the purely backwardlooking way of allocating capital that would characterize a capitalist economy with the predominantly passive style of investment. To the extent that the planning process would be forward looking and seek to understand the dynamics and future path of the development of the real economy, it could, we think, achieve a superior outcome compared with a system where all investment decisions are based on where capital has been allocated to in the past, as under a passive regime. And it is here that we think that individual policy makers have a vested interested in recognising and addressing the implicit cost-benefit tradeoffs that are implicit in the rise of passive management rather than regarding it as simply an unalloyed benefit. It seems clear to us that the rise of political disenchantment we have witnessed in the Western World cannot be separated from the economic backdrop of slower growth and rising inequality. Under such a situation, the promise of free market economies to deliver superior economic performance will increasingly be called into question and with it the mandate of the existing political order. It is at just such a time – where capital allocation in the real economy is the most challenging – that the efficiency of the financial economy is most sorely needed. Indeed it is at just such a time that the price of inefficiency, such as we would argue that the rise of passive management must induce, comes with the greatest cost. We show later in this report that active investing, by seeking to understand ex ante what the 'fair value' price of an asset, or an equilibrium price level for an industry is, and allocating capital accordingly, helps the process of price discovery to occur much faster than would otherwise be the case. This has clear social and economic benefits compared with a passive regime where capital flows at best do not help, and indeed can hinder, the price discovery process. We would argue that, by virtue of being forward looking, a process of planning of capital allocation in a Marxist society could by similar logic be superior to a largely passive regime where the capital allocation is done by a marginal participant based on past performance and without any regard to industry dynamics or deviations from fair value. Whether or not any planning process can 'beat' fully functioning capital markets with a meaningful share of AUM run actively, we can envisage such a process being more effective than largely passive capital markets at allocating capital- and so a Marxist regime being superior to a capitalist system with little or no active management. While this might seem a somewhat abstract debate, as one example, there is a very explicit point of relevance for policy on this point which is the Capital Markets Union (CMU) initiative of the EU, but it is by no means confined to Europe for its importance for policy. The CMU has the overarching objective of "maximising the benefits of capital markets and non-bank financial institutions for the real economy"5. The foundation of this initiative is a view expressed by policy makers that efficient capital markets are central to funding both growth and job creation. In the words of the policy document itself: "The CMU will help promote growth and financial stability. By facilitating companies' access to finance, in particular SMEs, the CMU will support growth and jobs' creation. At the same time, by promoting more diversified funding channels to the economy, it will help address possible risks stemming from the over-reliance on bank lending and intermediation in the financial system. By diversifying the risks, it will make the whole system more stable and help financial intermediaries granting more funding to the economy …. Overall, capital markets (especially equity markets) facilitate entrepreneurial and other risk-taking activities, which have a positive effect on economic growth. … Large and well-integrated capital markets can contribute to jobs and growth through a number of channels ".6

5 6

http://ec.europa.eu/finance/capital-markets-union/index_en.htm#action-plan ibid

FUND MANAGEMENT STRATEGY

BERNSTEIN

7

23 August 2016

EXHIBIT 3: A stylised view of the economic benefits of integrated and well-functioning capital markets

Source: Action Plan on building a Capital Markets Union - Accompanying working document: Economic analysis, European Commission, 30 September 2015

There is no explicit discussion of active vs passive in the commission staff working document. However, we think there should be. The link is in empirical academic work on the efficiency of capital markets and their impact on the allocation of capital. There is a small academic and policy literature on the role of equity markets in the capital allocation process in the economy but, to our knowledge, this has not so far been explicitly linked to the growth of passive investment, we attempt to make that link here. Wurgler (2000)7 finds evidence that equity markets do have an impact on the allocation of capital. What is of particular relevance here is that he shows that if the correlation of stocks increases then that impedes the efficient allocation of capital (the other explanatory variables that he finds important are minority shareholder rights and the extent of state participation). Wurgler uses a measure of the efficiency of the allocation of capital that measures whether there is increasing investment in industries that are growing and vice versa. Specifically, he measures the elasticity of gross fixed capital formation (essentially capex) to growth in profits across industries. The question that is posed is can the difference in this elasticity between countries be related to metrics of the development of financial markets. Ie do the financial markets in some countries do a better job of allocating capital to the optimal industries than others? The answer is a robust "yes". He concludes this analysis with: "Stock markets, particularly those that exhibit a high proportion of firm-specific price movements, appear to provide useful public signals of investment opportunities". The question of active vs passive is not discussed in the paper (the paper was published in 2000 when the share of AUM that was passive was trivially small). However, we can indirectly link this to the topic in hand. If there is evidence that passive investing increases correlation – be that a temporary shock to correlation or a more sustained change – then presumably there is the possibility that this could in turn affect the elasticity of capital investment to optimal industries. In our own research we have taken a similar line. We first note that finding a simple short-term link between passive asset share and correlation is difficult as passive asset share has risen monotonically for the past decade while average pairwise stock correlation is cyclical, Exhibit 4, (though in the recent work of Bolla, Kohler and Wittig (2016), which we discuss below, they do find such a relationship). But let us think through the process at work here in a stylised way. Consider a market that undergoes an increase in the proportion of assets managed passively from one period to the next. In the second period more securities will be trading in line with a macro view (ie how they are priced with regard to an index or ETF product, say) rather than with respect to "fundamentals". Therefore, over the course of the transition from period 1 to period 2 the average pairwise correlation of the securities would have increased. But that means that in period 2 there will always be a small group of active managers who will spot the mis-pricing of securities and put trades that will eventually nudge the securities back to their fundamental level

7

Wurgler (2000): Financial markets and the allocation of capital

FUND MANAGEMENT STRATEGY

BERNSTEIN

8

23 August 2016

accordingly. The result of all this would be a spike in correlation that eventually mean-reverts. It would imply that when such a spike occurs that it is larger if the passive asset share of that market is larger8. Is there any evidence of this happening in practice? We think that there is. Rather than looking at the long run relationship between average pairwise stock correlation and passive asset share we focus on the peak correlation reached over the cycle. One difficulty with empirically testing this approach is that although there have been very significant spikes in correlation over the last decade there have not been many of them. To increase the breadth of the data set we partition it by region as the passive asset share has persistently been higher in the US than in Europe at the point of each spike. We also partition between large cap and small cap stocks because again the large cap stocks have higher passive penetration. The result is Exhibit 5 where we plot the maximum correlation reached at each point for securities in each regional/size segment against the passive asset share of fund AUM for that segment at that point in time. The positive line of best fit is at least prima facie evidence that a higher passive asset share leads to greater spikes in correlation between stocks.

EXHIBIT 4: No simple relationship between average stock correlation and passive share 40 35

EXHIBIT 5: … but higher passive share is associated with higher correlation spikes

0.30

0.8

0.25

0.7

y = 0.0042x + 0.3638 R² = 0.4161

%

20

0.15

15

0.10

10 0.05

5

0.00

Jan-16

Jan-14

Jan-12

Jan-10

Jan-08

Jan-06

Jan-04

Jan-02

Jan-00

0

Cross-Sectional Correlation

0.20

25

Correlation

30

0.6 0.5 0.4

Large cap

0.3

Small cap

0.2 0.1

Global: Passive Share (LHS)

0.0 MSCI AC World Global Pairwise Correlation with 180-Day Lookback Window (RHS)

0

20

40

60

80

Passive Share Source: EPFR, MSCI, Bernstein analysis

The scatter plot shows the peaks in cross-sectional correlation for US and Europe large and small cap styles in the periods (2003, 2007/08, 2009, 2010, 2011, 2015) against their respective level of passive share. To construct the large and small cap passive share series we adjusted the passive share of US and Europe broad market indices by the ratio of passive share for S&P 500 vs Russell 2000. Source: eVestment, EPFR, S&P, Russell, MSCI, Bernstein analysis

9 Continuing this theme, Levine and Zervos (1998) show that stock market liquidity and banking development are both predictors of growth, capital accumulation, and productivity improvements. They conclude that this provides evidence that stock markets are important for economic growth in a way that is distinct from the provision of credit supplied by banks. This

8

Note that we are being purposely vague here about the time scale and whether this process more naturally takes place over days, months or years. 9 Levine and Zervos (1998): Stock Markets, Banks, and Economic Growth: Do well-functioning stock markets and banks promote long-run economic growth? worldbank.org

FUND MANAGEMENT STRATEGY

BERNSTEIN

9

23 August 2016

paper was also cited by the SEC as an example of why a well-functioning capital markets are important for macroeconomic growth10. 11 Taking a different angle, Rajan and Zingales (1996) recognise that some industries require external financing for capital intensive projects more than others. So they specifically focus on industries that are more dependent on external finance and conclude that they grow faster in countries that have more developed financial markets. They determine which industries are more in need of external finance from the difference between investments and internal cash from operations. They then analyse whether such industries grow faster in countries that have better-developed financial markets. They find that:

In a country which is one standard deviation above the mean of financial development, the difference in growth rates between an industry whose financial dependence [on external capital] is one standard deviation above the mean and the average industry, is 1% more (annually, and in real terms) than in the average country … [and] the intensity of investment in industries dependent on external finance is disproportionately higher in countries with more developed financial markets12. Their definition of financial development for a country is market cap/GDP, domestic credit/GDP and the ratio of credit raised by the non-financial private sector as a proportion of total domestic credit. Again, this analysis was conducted long before passive management was a topic of concern and so this measure of financial development cannot be related to the proportion of the market that is passively managed. However, it is instead evidence of the broader need for a well-functioning capital market for capital-intensive industries. An intriguing suggestion has been made13 that more passive allocation might even improve capital allocation by removing less skilled asset managers. If it was true that any incremental flow out of active into passive did indeed first take away funds from managers who were less good at effective capital allocation then there may be a case for this hypothesis. But it is notoriously hard to identify outperforming managers ex ante and many frictional inefficiencies in the allocation towards managers also suggest that in practice it is unlikely that it is always the worst manager who loses assets in any marginal allocation to passive. If, as is more likely, it is a manager with a random level of skill who loses out then this argument would not hold and in fact there would be a reduction in the AUM of potentially skilled asset allocation. The fundamental problem with the view that passive allocation can act in this manner is that it implicitly assumes that an unambiguous criterion of success can be established. However the issue is that any measure of success is time dependent (over which horizon should we measure performance, daily, monthly, yearly…?) and herein lies the problem. There is a fundamental time scale over which total system economic value created is determined by the features of the real economy. It is dictated by the payback period on investment in the real economy which is itself determined by the lead time over which capital is deployed (which can be several years or more), the capital intensity of the industry and its profitability. In the real economy this fundamental time horizon has been getting ever more extended – the mining sector, which we address in detail, perfectly illustrates this tendency. To the extent that the financial economy serves as a forward looking mechanism for price stabilisation and mean reversion, then the horizon over which success is determined and measured must extend commensurately. However, the aggregate time horizon of the market as a whole has become foreshortened rather than extended by the rise of passive management (we discuss this in more detail at the end of this note). By definition, passive flows of capital, given that they seek to emulate or replicate what has already occurred must be backward looking. Moreover if the aggregate time horizon discounted by the market is the sum of the forward looking horizon of active management and the negative backward looking horizon of passive management then the window of visibility has been foreshortened. Consequently the measurement of investment success, against which flows of capital are supposed to be allocated within the financial economy, is moving against the trend that is required to justify the social utility of the financial markets.

10

Aguilar (2015): U.S. Equity Market Structure: Making Our Markets Work Better for Investors, SEC Public Statement. May 11, 2015. See note 39 in the SEC document. 11 Rajan and Zingales (1996): Financial dependence and Growth, NBER working paper 5758 available at http://www.nber.org/papers/w5758 12 ibid 13 Philosophical Economics (2016a)

FUND MANAGEMENT STRATEGY

BERNSTEIN

10

23 August 2016

MARXISM, THE MARKETS & MINING – A VIEW FROM THE REAL ECONOMY. Given that we have introduced the contrast between Marxist and market models of capital allocation, perhaps the best explanation of why then we digress into mining can be given by Marx himself. As the opening of Das Kapital puts it;

"The wealth of those societies in which the capitalist mode of production prevails presents itself as an immense accumulation of commodities" Fundamentally this note is a discussion about the nature of the connection between the financial and the real economy. Ultimately it is a disquisition into the purpose of the real economy and how the financial economy aids or abets the realisation of that purpose (as Aristotle would put it, it is an analysis into both the final and efficient causes of economic growth). In this regard, mining is an activity in the real economy fundamentally very well suited to the current discussion. It stands both as the most capital intensive of all activities in the developed world and also as the ultimate origin of all the physical capital that, via the process of capital accumulation highlighted by Marx, stands as the basis of the wealth of nations. We take it to be almost axiomatically true that the purpose of the financial economy is to aid capital allocation decisions in the real economy. To the extent that this is not the case then any financial market is simply a zero sum game that is capable of redistributing the calls on the productive capacity of the real economy but has no role in changing that capacity. It may be argued that the redistributive effects of the financial market serve some other social purpose and it is this alternative that establishes the warrant for their existence, but this seems tenuous at best. Moreover, given the existence of frictional costs in trading, asymmetries in knowledge and other well know market imperfections the purely redistributive nature of the markets will always have a tendency to lead to less rather than more socially acceptable outcomes. However, to the extent that financial markets aid capital allocation in the real economy then there is an economic case to be made for their existence. The superiority of capital allocation in the real economy given the presence of the financial economy justifies the calls that those engaged in the financial markets make upon the production of the real economy. Of course this leads immediately to the requirement to define what exactly we mean by "capital allocation" and indeed how we would measure a superior versus inferior model of capital allocation. The capital any economy possesses is ultimately a reflection, or crystallisation of its savings and the excess of current production above current consumptive requirements. Those savings need not be invested but simply hoarded, and of course the result of that is that neither overall consumption nor production are thereby increased but simply smoothed or averaged out. Of course saving and capital in this sense is as old as humanity (as the quote below from Genesis illustrates) but does not really help understand the process of capital allocation per se.

"And Joseph went out from the presence of Pharaoh, and went throughout all the land of Egypt. Now in the seven plentiful years the ground brought forth abundantly. So he gathered up all the food of the seven years which were in the land of Egypt, and laid up the food in the cities; he laid up in every city the food of the fields which surrounded them. Joseph gathered very much grain, as the sand of the sea, until he stopped counting, for it was immeasurable…Then the seven years of plenty which were in the land of Egypt ended, and the seven years of famine began to come, as Joseph had said. The famine was in all lands, but in all the land of Egypt there was bread. So when all the land of Egypt was famished, the people cried to Pharaoh for bread. Then Pharaoh said to all the Egyptians, 'Go to Joseph'." However, to the extent that those savings are invested the clear intent is that there should be a magnification of the consumption forgone in the present so as to deliver increased consumption in the future. Growth in overall economic output follows on the repeated application of saving a portion of current output in investments capable of this multiplication. A fairly obvious corollary of this is that it is possible to disaggregate the trend growth in output that should result from such a model into two components, namely the savings rate in the economy and the internal rate of return on the investment to which those savings are allocated. Thus a savings rate of 40% into investments yielding a 10% IRR will generate trend economic growth of 4%. The growth in any economy will be maximised when the available savings are allocated to the highest returning investment opportunities. What differentiates the modern from the premodern economy is, in part, the availability of investment opportunities whose IRR is sufficiently great that relatively modest savings rates are capable of generating prodigious rates of growth in overall production. Of course the availability of the set of those IRR accretive investment opportunities is primarily the result of technological innovation, but technological innovation in itself is insufficient to engender growth. After all it could be argued that the principles of the steam engine were latent in the Aeolipile of Hero of Alexandria in the first century AD. What is required as

FUND MANAGEMENT STRATEGY

BERNSTEIN

11

23 August 2016

much as the innovation is the social technology that allows for its realisation. The scientific principles required for a very significant industrial revolution were known in antiquity and yet were left unrealised. The impediment was the essentially two fold; in the first place the abundance of free human labour in the form of slaves undermined the economic incentives to innovate and secondly the social conservatism of a period which was more concerned about protecting the existing status quo than "progress". One is immediately reminded of the passage in De lapidibus et metallis from book 16 of the Etymologies of St Isidore of Seville written in the early 600s recounting the reign of the Emperor Tiberius.

"Under Tiberius Caesar a certain craftsman devised a formula for glass so that it would be flexible and pliable. And when he was brought before Caesar he presented a drinking bowl to him, and Caesar indignantly threw it to the floor. The craftsman picked the drinking bowl up from the floor, where it had been dented as a bronze vessel would be. Then he took a small hammer from his pocket and reshaped the drinking bowl. When he had done this, Caesar said to him, “Does anyone else know this method of making glass?” After the craftsman swore that no one else knew, Caesar ordered him beheaded, lest, if this skill became known, gold would be regarded as mud and the value of all metals would be reduced – and it is true that if glass vessels became unbreakable, they would be better than gold and silver." The very simple schematic model of economic growth that captures these two critical features is adapted from that of the Austrian economist Mark Skousen in his book "The Structure of Production" and is illustrated below.

EXHIBIT 6: Our simple "steady state" model of the economy. Savings, investment and the return (IRR) on investment drive the creation of a stock of consumer goods whose use or "consumption" delivers utility

𝐾𝐾𝑡𝑡+1 = 𝐾𝐾𝑡𝑡 ∙ �1 −

1 � + 𝐼𝐼𝑡𝑡 𝑎𝑎𝑎𝑎𝑎𝑎 𝐾𝐾0 = 𝑎𝑎 𝐿𝐿𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶

𝑃𝑃𝑡𝑡 = 𝐼𝐼𝐼𝐼𝐼𝐼 ∙ 𝐾𝐾𝑡𝑡

A certain fraction of the output (P) of the world's capital stock goes to produce further capital goods, i.e. it is saved in investment (I) while a proportion goes to the production of consumer goods (A) of varying degrees of durability and hence useful life. Thus; 𝐼𝐼𝑡𝑡 = 𝑆𝑆𝑡𝑡 ∙ 𝑃𝑃𝑡𝑡

𝐴𝐴𝑡𝑡 = (1 − 𝑆𝑆𝑡𝑡 ) ∙ 𝑃𝑃𝑡𝑡

Thus the stock of consumer goods (Q – and by definition the matter used to support those goods) grows at a corresponding rate (where here R captures the recycle rate of any raw material) 𝑄𝑄𝑡𝑡+1 = 𝑄𝑄𝑡𝑡 ∙ �1 −

(1 − 𝑅𝑅) � + 𝐴𝐴𝑡𝑡 and 𝑄𝑄0 = 𝑏𝑏 𝐿𝐿𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶

The stock of durable consumer goods is then "consumed" at a rate which defines the velocity of goods in the economy. 𝐶𝐶𝑡𝑡 = 𝑄𝑄𝑡𝑡 ∙ 𝑉𝑉𝑡𝑡

The total output being the sum of investment and consumption (absent import-export corrections). 𝐺𝐺𝐺𝐺𝐺𝐺𝑡𝑡 = 𝐼𝐼𝑡𝑡 + 𝐶𝐶𝑡𝑡 = 𝐼𝐼𝑡𝑡 + 𝑄𝑄𝑡𝑡 ∙ 𝑉𝑉𝑡𝑡

The implications of the above are, clearly, that trend growth rate in any economy – assuming a constant velocity of goods – is the product of the savings rate and the IRR at which those savings are invested. lim % 𝐺𝐺𝐺𝐺𝐺𝐺 = %𝐼𝐼𝐼𝐼𝐼𝐼 ∙ %𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆

𝑡𝑡→∞

Source: Bernstein Analysis

FUND MANAGEMENT STRATEGY

BERNSTEIN

12

23 August 2016

EXHIBIT 7: A simple model of the economy…savings drives investment, the return on which delivers growth Utility

Primary Production

Transformation (Secondary)

Stock of Consumer Goods

Consumer Goods

Production

+

Capital Goods

Source: Bernstein Analysis

EXHIBIT 8: Trend growth rates determined by the product of the savings rate and the IRR on investment. To create growth an economy must either save more or have access to technological innovation capable of replenishing the stock of IRR accretive investment opportunities. Model Predictions - GDP Growth Rate 7.0%

GDP Growth Rate

6.0% Trend growth rate = IRR on Investment * Savings Rate

5.0% 4.0% 3.0% 2.0% 1.0% 0.0% 1

3

5

7

9

11

13

15

17

19

21

23

25

27

29

31

33

35

37

39

41

43

45

47

49

Years Trend Growth Rate

YoY Growth Rate

Source: Bernstein Analysis

Now while all this appears to be somewhat "economics 101" it seem important because unless one has a definite model of how a the real economy actually works and a clear definition of what is meant by capital within that model any discussion of the role of the financial economy will simply be a non-starter. It is only with reference to such a model that we can even hope to make

FUND MANAGEMENT STRATEGY

BERNSTEIN

13

23 August 2016

sense of any subsequent discussion about the outcomes of different models of capital allocation. Furthermore, under this frame work it is possible to illustrate quite powerfully the progress over time of different economic models; in particular between the free-market models to the West and the command & control economy of the Soviet Union.

EXHIBIT 9: Showing the history of the Soviet Union, the USA and Germany since 1900…upward sloping lines indicate investment in IRR positive projects, downward sloping lines the opposite. Paradigms for Embedding Steel Stock 50

Output — GDP per Capita ($1,000)

45 40 35 30 25 20 15 10 5 0 0

2,000

4,000

6,000

8,000

10,000

12,000

14,000

16,000

18,000

20,000

22,000

24,000

26,000

28,000

30,000

32,000

Capital Stock — kg Steel per Capita USA USSR Germany

Source: IMF, Maddison, Bernstein Analysis

FUND MANAGEMENT STRATEGY

BERNSTEIN

14

23 August 2016

EXHIBIT 10: The Transpolar Railway between Salekhard and Nadym, which was abandoned after Stalin's death, serves as a case in point.

Source: Wikimedia Commons

Now the reason why any mining analyst should be interested in the various models of capital allocation and the impact that the financial economy has on this should be obvious given the cost structure of the mining industry. Mining is the most capital intensive activity of the modern economy. While this was not always the case, the progress in productivity in this industry has, over the last 100 years, been synonymous with the elimination of labour as a factor of production. It is a similar story in all the other primary extractive industries. It is these sectors of the economy that require capital and it is in these sectors of the economy more than anywhere else that efficient capital allocation is paramount.

FUND MANAGEMENT STRATEGY

BERNSTEIN

15

23 August 2016

EXHIBIT 11: The increase in mining productivity has been driven by the intensive capitalization of the mining industry. Mining is the most capital intensive industry in modern society. Factor Share of Gross Output by Sector 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Capital cost shares

Intermediate inputs cost shares

Labour cost shares

Source: ABS, Bernstein Analysis & Estimates

EXHIBIT 12: Labour accounts for just 20% of the factor input into mining, compared to an average of over 60% in the wider economy. Factor Share of Value Add by Sector 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Capital cost shares

Labour cost shares

Source: ABS, Bernstein Analysis & Estimates

FUND MANAGEMENT STRATEGY

BERNSTEIN

16

23 August 2016

EXHIBIT 13: The structural progression of mining capitalization has been remarkable. Long Term Factor Share of Mining Output 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1870

1900

1950 Labour

2000

Capital

Source: Kendrick, ABS, Bernstein Analysis & Estimates

Now it is perfectly possible to conceive of the real economy functioning without the supporting action of the financial economy. Considering the development of the real economy under this counterfactual is a useful way of highlighting the benefit created by the financial economy. In the first place we can assert an initial quantity for any particular good or service with the following supply and demand equations 𝐷𝐷𝑡𝑡 = 𝑎𝑎 − 𝑏𝑏 ∙ 𝑃𝑃𝑡𝑡 𝑛𝑛

𝑆𝑆𝑡𝑡 = ��

(𝑃𝑃𝑡𝑡 − 𝛼𝛼) � 𝛽𝛽

These two equations are simply an expression for an inflected cost curve intersecting a backward sloping demand line, as shown below. Given that we are also using a non-linear and inflected cost curve in this analysis we also provide some supplemental analysis of how to interpret the variables that we use in the supply expression above.

FUND MANAGEMENT STRATEGY

BERNSTEIN

17

23 August 2016

EXHIBIT 14: Price set so as to balance supply and demand. Illustrative Example of Price Discovery 70.0 60.0

Demand = a - b*P a is the intercept of the demand curve, the price at which demand is sent to zero, b is the slope of the demand curve

Cost/Price

50.0 40.0

Supply, given by three parameters, α, β and n α is the cost of the lowest cost producer, β is the steepness of the cost curve and n the degree of inflection in the curve.

30.0 20.0 10.0 0.0 0%

20%

10%

30%

40%

50%

60%

70%

80%

90%

100%

Dumulative Supply/Demand Supply

Demand

Source: Bernstein Analysis

EXHIBIT 15: Impact of changing α in the EXHIBIT 16: Impact of changing β in the EXHIBIT 17: Impact of changing n in the cost curve. cost curve. cost curve.

60 50

70

Cost of Production - $/t

Cost of Production - $/t

70

40 30 20 10

50

40

40 30 20 10

0%

20%

40%

60%

80%

Cumulative Supply

Source: Bernstein Analysis

100%

35 30 25 20 15 10 5

0

0

Higher price level and margin due to inflection of cost curve but supported by fewer high cost units

45

Higher price level and margin due to steepness of cost curve

60

Cost of Production

Constant level of margin generated even though prices will be different

80

Example Cost Curve - Inflection

Example Cost Curve - Steepness

Example Cost Curve - Level

0 0%

20%

40%

60%

80%

100%

0%

Cumulative Supply

Source: Bernstein Analysis

20%

40%

60%

80%

100%

Cumulative Supply

Source: Bernstein Analysis

The action of the market is to ensure that supply and demand equal each other through the action of price. Now on the supply side we have an inflected cost curve and this inflection represents the existence of the steeping and flattening of cost structures that give rise to the differences in short and long term elasticities of supply. Thus we identify the equilibrium supply and demand at any point in time as Q. 𝐷𝐷𝑡𝑡 = 𝑆𝑆𝑡𝑡 = 𝑄𝑄𝑡𝑡

Under the condition of price equilibrium described above we can model the financial position of the industry as follows. 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 = 𝑄𝑄𝑡𝑡 ∙ 𝑃𝑃𝑡𝑡 𝑄𝑄𝑡𝑡

𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝑜𝑜𝑜𝑜 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 = � (𝛼𝛼 + 𝛽𝛽 ∙ 𝑄𝑄 𝑛𝑛 ) 𝑑𝑑𝑑𝑑 = 𝛼𝛼 ∙ 𝑄𝑄𝑡𝑡 + 0

FUND MANAGEMENT STRATEGY

𝛽𝛽 ∙ 𝑄𝑄 𝑛𝑛+1 𝑛𝑛 + 1 𝑡𝑡

BERNSTEIN

18

23 August 2016

𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 = 𝑄𝑄𝑡𝑡 ∙ 𝑃𝑃𝑡𝑡 − 𝛼𝛼 ∙ 𝑄𝑄𝑡𝑡 −

𝛽𝛽 𝛽𝛽 ∙ 𝑛𝑛 𝑛𝑛 ∙ 𝑄𝑄 𝑛𝑛+1 = ∙ 𝑄𝑄 𝑛𝑛+1 = (𝑃𝑃𝑡𝑡 − 𝛼𝛼) ∙ 𝑄𝑄𝑡𝑡 ∙ 𝑛𝑛 + 1 𝑡𝑡 𝑛𝑛 + 1 𝑡𝑡 𝑛𝑛 + 1

Now assuming that there is a capital intensity of production of K then at any point in time the installed capital base will be KQ and given a depreciation rate d then the depreciation will be KQd and thus 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 = (𝑃𝑃𝑡𝑡 − 𝛼𝛼) ∙ 𝑄𝑄𝑡𝑡 ∙

𝑛𝑛 − 𝐾𝐾 ∙ 𝑄𝑄𝑡𝑡 ∙ 𝑑𝑑 𝑛𝑛 + 1

The identities establish the margins and the returns on capital (ROCE) that characterise the industry at any point in time. Furthermore at a tax rate of T the tax paid by the industry will therefore be 𝑇𝑇𝑇𝑇𝑇𝑇 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 = 𝑇𝑇 ∙ �(𝑃𝑃𝑡𝑡 − 𝛼𝛼) ∙ 𝑄𝑄𝑡𝑡 ∙

And

𝑛𝑛 − 𝐾𝐾 ∙ 𝑄𝑄𝑡𝑡 ∙ 𝑑𝑑 � 𝑛𝑛 + 1

𝑃𝑃𝑃𝑃𝑃𝑃 = 𝑄𝑄𝑡𝑡 ∙ (1 − 𝑇𝑇) ∙ �(𝑃𝑃𝑡𝑡 − 𝛼𝛼) ∙

𝑛𝑛 − 𝐾𝐾 ∙ 𝑑𝑑� 𝑛𝑛 + 1

Given a dividend pay-out ratio of D the dividend paid out by the industry will be, 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 = 𝐷𝐷 ∙ 𝑄𝑄𝑡𝑡 ∙ (1 − 𝑇𝑇) ∙ �(𝑃𝑃𝑡𝑡 − 𝛼𝛼) ∙

𝑛𝑛 − 𝐾𝐾 ∙ 𝑑𝑑� 𝑛𝑛 + 1

Thus the residual cash generated by the industry and available for reinvestment will be

𝑅𝑅𝑅𝑅𝑅𝑅 =

𝑅𝑅𝑅𝑅𝑅𝑅 = 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 − 𝑇𝑇𝑇𝑇𝑇𝑇 − 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷

(𝑃𝑃𝑡𝑡 − 𝛼𝛼) ∙ 𝑄𝑄𝑡𝑡 ∙ 𝑛𝑛 ∙ (1 − 𝑇𝑇) ∙ (1 − 𝐷𝐷) + 𝐾𝐾 ∙ 𝑄𝑄𝑡𝑡 ∙ 𝑑𝑑 ∙ [(1 − 𝑇𝑇) ∙ 𝐷𝐷 + 𝑇𝑇] 𝑛𝑛 + 1

This residual cash in invested in capital projects with an average capital intensity K and so will generate an increase in mine supply after a number of years equal to the average lead time for new production (for example it takes between 3 and 6 years to build a new mine). ∆𝑄𝑄 =

𝑅𝑅𝑅𝑅𝑅𝑅 𝐾𝐾

We then have a situation where there is an increase in supply that will change the structure of the supply equations at some subsequent point in the future as a result of the present profitability of the industry. The classic example of this is the cost curve flattening that takes place in mining following periods of elevated prices. The elevated margins suck a huge amount of capital into the industry, that capital is invested in low cost growth projects which come to the market many years after the initial investment was made (again, say, between 3 to 6 years later) given the long construction lead times. However, as the projects come on line they change the structure of the industry and act to flatten cost curves and lower overall profitability. Now there are any number of ways that one can model this cost curve transformation, one easy way that corresponds to an aggregate flattening of curves through the introduction of low cost supply would be as follows. 𝑛𝑛𝑡𝑡+1 = 𝑛𝑛𝑡𝑡 ∙

𝑙𝑙𝑙𝑙(𝑄𝑄𝑡𝑡 ) 𝑙𝑙𝑙𝑙(𝑄𝑄𝑡𝑡+1 )

Another would be through an assumption that new supply enters the industry evenly across the cost curve. Under this assumption the transformation would be as follows. 𝛽𝛽𝑡𝑡+1 = 𝛽𝛽𝑡𝑡 ∙ �

𝑄𝑄𝑡𝑡 𝑛𝑛 � 𝑄𝑄𝑡𝑡+1

At the same time the change in volume created through the action of investment in new projects offsetting depletion is given by. 𝑄𝑄𝑡𝑡 = 𝑄𝑄𝑡𝑡−1 ∙ (1 − 𝑑𝑑) + ∆𝑄𝑄𝑡𝑡−𝜏𝜏

In other words the supply at any given point in time is equal to the previous year's supply less depletion plus new capacity commissioned as a consequence of investment made a number of years (τ) previously.

FUND MANAGEMENT STRATEGY

BERNSTEIN

19

23 August 2016

We can also represent the emergence of demand growth, g, for any industry by changing the parameters of the demand equation over time as follows. 𝑎𝑎𝑡𝑡+1 = 𝑎𝑎𝑡𝑡 ∙ (1 + 𝑔𝑔)

Under this assumption we have demand growth occurring through an increase in the number of consumers each with identical consumptive preferences for the utility engendered by possession of a certain quantity of the good in question. To the extent that the utility of the good itself changes or new goods are introduced then the second parameter (b) in the demand equation can be altered to reflect this. Now, on the other side of the equation, for there to be price equilibrium, the cost curve must be such that the margin being generated by the industry covers the depletion rate of the existing asset base plus the desired growth rate. Given that each incremental unit of supply comes with a capital cost of K the total cash flow that the industry needs to retain and reinvest is given by the following. 𝑅𝑅𝑅𝑅𝑅𝑅 = 𝐾𝐾 ∙ 𝑄𝑄𝑡𝑡 ∙ (𝑔𝑔 + 𝑑𝑑)

Where g is the growth in demand and d is the depletion rate as per the previous discussion. To the extent that the industry generates a cash flow greater than this equilibrium level then the growth in low cost supply will continue to flatten the cost curve. This cost curve flattening will only come to an end once the equilibrium condition is reached. However, it should also be remembered that there is a lead time between investment and new supply appearing. Consequently the cash retained at any point in time must be equal to the equilibrium supply condition – i.e. growth in supply equalling depletion plus demand – a number of years into the future. Adjusting for this gives the following expression. 𝑅𝑅𝑅𝑅𝑅𝑅 = 𝐾𝐾 ∙ 𝑄𝑄𝑡𝑡 ∙ (1 + 𝑔𝑔)𝜏𝜏 ∙ (𝑔𝑔 + 𝑑𝑑)

Thus we can establish the equilibrium price for any commodity by the identity between the two expressions for the retained cash flow of the industry. (𝑃𝑃𝑡𝑡 − 𝛼𝛼) ∙ 𝑄𝑄𝑡𝑡 ∙ 𝑛𝑛 ∙ (1 − 𝑇𝑇) ∙ (1 − 𝐷𝐷) + 𝐾𝐾 ∙ 𝑄𝑄𝑡𝑡 ∙ 𝑑𝑑 ∙ [(1 − 𝑇𝑇) ∙ 𝐷𝐷 + 𝑇𝑇] = 𝐾𝐾 ∙ 𝑄𝑄𝑡𝑡 ∙ (1 + 𝑔𝑔)𝜏𝜏 ∙ (𝑔𝑔 + 𝑑𝑑) 𝑛𝑛 + 1

Reassembling the above gives the equilibrium price condition 𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 = 𝛼𝛼 +

𝐾𝐾 ∙ (𝑛𝑛 + 1) ∙ �(1 + 𝑔𝑔)𝜏𝜏 ∙ (𝑔𝑔 + 𝑑𝑑) − 𝑑𝑑 ∙ [(1 − 𝑇𝑇) ∙ 𝐷𝐷 + 𝑇𝑇]� 𝑛𝑛 ∙ (1 − 𝑇𝑇) ∙ (1 − 𝐷𝐷)

The iterative set of equations described above can then be solved at each point in time to generate a self-consistent picture of the development of an industry under its own financial resources without the intervention of the financial economy. Thus we have a condition for the long term price of any commodity entirely determined by the economic structure of the industry. 

The costs of the low cost producer of the commodity (α)



The shape of the cost curve (n)



The capital intensity of the industry (K)



The growth in demand (g)



The depletion rate (d)



The lead time for new build on investments (τ)



The tax rate (T)



The dividend pay-out ratio for the industry (D)

Moreover we can see that the equilibrium EBITDA and EBIT margins for the industry emerge quite naturally from this framework as does the equilibrium return on capital (ROCE).

FUND MANAGEMENT STRATEGY

BERNSTEIN

20

23 August 2016

𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 = 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 = 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 =

�𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒 − 𝛼𝛼� 𝑛𝑛 ∙ 𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛 + 1

�𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒 − 𝛼𝛼� 𝑛𝑛 𝐾𝐾 ∙ 𝑑𝑑 ∙ − 𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛 + 1 𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒

�𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒 − 𝛼𝛼� 𝑛𝑛 ∙ − 𝑑𝑑 𝐾𝐾 𝑛𝑛 + 1

Any commodity industry will automatically converge to this price level and level of financial performance as a matter of course over time. Short term dis-equilibrating shocks to the industry are quickly dissipated away and the long run price equilibrium will always reassert itself. While at every point in time the price of the commodity is set by the marginal cash costs of production this approach explicitly recognises that these marginal cash costs are themselves a function of the reinvestment rate of the entire industry and the steepening or flattening of the costs of production that result from this process. A useful way of summarising this result is to say that a genuine price equilibrium is reached when the price level implied by the marginal cash costs of production is equal to the price level implied by the lowest cost producer earning a full return on capital investment. So what we see is a price formulation that integrates both the insights of incentive pricin g methodologies and cost curve methodologies. Commodity prices will continue to fall to the level dictated by the return on capital the shareholders of the lowest cost producer in any industry. This in turn is dictated by the dividend that these shareholders demand. The dividend required by these shareholders determines the steady state inflow of capital into the industry and thereby the rate of low cost reinvestment and the rate of the flattening of the cost curve. We can see this most clearly if we reverse the equation given for the industry ROCE. 𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒 = 𝛼𝛼 + 𝐾𝐾 ∙ (𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 + 𝑑𝑑) ∙

𝑛𝑛 + 1 𝑛𝑛

Under this we see the equilibrium price level is determined not so much by a focus on the marginal, i.e. high cost producer, but rather by the low cost producer (whose costs we denote as α), the capital intensity of production K and the average industry return on capital gross of underlying depletion. The final factor in the equation above (i.e. containing the terms n above) simply reflects the averaging of returns across the industry as a whole versus those enjoyed by the low cost producer. To consider what this means in the concrete, it might be worth giving the following example 

The costs of the low cost producer of the commodity (α) – US$20/t



The shape of the cost curve (n) – 1 (i.e. an linearly upward sloping curve)



The capital intensity of the industry (K) – US$150/t



The growth in demand (g) – 2.5%



The depletion rate (d) – 5% (i.e. on average 20 year life of mines)



The lead time for new build on investments (τ) – 4 years



The tax rate (T) – 35%



The dividend pay-out ratio for the industry (D) – 40%

If we put these parameters into the equilibrium price and margin expressions above the resulting equilibrium price for the industry is US$60/t, the average industry EBITDA margin at equilibrium will be 33% and the EBIT margin will be 21% while the industry average ROCE will be 8.4%. The other way to formulate this is to assert that the average industry return on capital is known ex ante and to this into the inverted ROCE formula given above to generate the corresponding equilibrium price (essentially saying that all industries are roughly cost of capital on average and deducing the price level consistent with this).

Which, in this case, is as follows;

FUND MANAGEMENT STRATEGY

𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒 = 𝛼𝛼 + 𝐾𝐾 ∙ (𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 + 𝑑𝑑) ∙

𝑛𝑛 + 1 𝑛𝑛

BERNSTEIN

21

23 August 2016

𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒 = 20 + 150 ∙ (8.4% + 5%) ∙

1+1 = 60 1

So the same equilibrium price is achieved and this corresponds to the lowest cost producer in the industry earning a 21.8% ROCE with the industry as a whole generating an 8.4% return. So another way of looking at this is to say that the equilibrium price for any commodity is set by this implied ROCE that the lowest cost producer in the industry ought to enjoy, with every other player that makes up the total industry average ROCE earning a lower return. Putting the equilibrium condition in these terms shows that what determines price in equilibrium is the economics of the lowest cost producer and not the marginal producer (even though at every point in time marginal supply and marginal demand are in an instantaneous equilibrium). Furthermore we can see exactly how the returns generated by the industry as a whole and the lowest cost producer a determined as a function of trend demand growth rates.

EXHIBIT 18: Return Sensitivity as Function of Industry Demand Growth

Growth Rate

Average ROCE

Low Cost ROCE

0.0%

0.0%

5.0%

0.5%

1.6%

8.1%

1.0%

3.2%

11.4%

1.5%

4.9%

14.7%

2.0%

6.6%

18.2%

2.5%

8.4%

21.8%

3.0%

10.3%

25.5%

3.5%

12.2%

29.4%

4.0%

14.2%

33.4%

Source: Bernstein Analysis & Estimates

In the economic model introduced at the start of this note we had trend growth in output determined by the product of the savings rate and the IRR at which those savings are reinvested. Now in the extension of this economic model to the micro level of an industry the meaning of the savings rate is pretty clear, it is the post-tax and post-dividend operating cash flow retained by the industry. The issue was that at a micro level the IRR on investment is an outcome of the investment process not an input into it. The inputs are the capital intensity of investment, the lead time for project development, the operating costs of production and so forth. However the extension to the model introduced here shows how this problem is overcome. The equilibrium ROCE (and hence the IRR) on the investment emerges as a consequence of the development of the equilibrium price level via the action that new investment has on the operating cost structure of the industry. Another feature of this analysis is that it highlights the importance of the dividend policy of an industry. One of the main issues with the dividend policy is that the majority of conventional valuation analysis treats dividends as being a value neutral proposition; after all dividends are cash that accrues to the beneficial owners of an industry whether or not it is formally returned to them. Under this view of the world it is simply a question of deciding in which pocket shareholders wish to keep their cash. However what we show explicitly here is that this is too simplistic a picture. It is the dividend policy of an industry that is one of the primary determinants of the equilibrium price level and hence the equilibrium ROCE that the industry generates. A high dividend pay-out removes excess liquidity from an industry meaning that the reinvestment rate of the industry will be lower. In turn this means, all other things being equal, that the price level at which the margin generated and retained by the industry is sufficient to meet demand growth will be higher. The lower the dividend the lower the price level and the lower the resulting ROCE that any industry generates, and vice versa. Consequently the dividend policy is far from being a value neutral measure but is the most important lever by which shareholders can actually influence the returns that they will enjoy. We show this sensitivity below.

FUND MANAGEMENT STRATEGY

BERNSTEIN

22

23 August 2016

Exhibit 19: Return Sensitivity as Function of Industry Dividend Policy

Dividend Payout Ratio

Industry ROCE

0.0%

5.0%

10.0%

5.6%

20.0%

6.3%

30.0%

7.2%

40.0%

8.4%

50.0%

10.1%

60.0%

12.6%

70.0%

16.8%

80.0%

25.2%

Source: Bernstein Analysis & Estimates

A further point that should be heavily stressed is that in the above equilibrium price condition what we show is that the price level is related to the growth rate of an economy. A flat real price line ought, in equilibrium, be able to generate a constant growth in output. But the implication of this is that it is changes in the growth rate should be proportional to changes in the price line. Thus it is the acceleration/deceleration in GDP that drives changes in commodity price levels. Thus it is no surprise that we see the extreme sensitivity of commodity prices to the overall macro environment. Now, as interesting as all this is, so far there has been no connection made to the role of the financial markets in the above industry model so what is the connection? The answer lies in part in the amount of time that it would take any industry to move to the equilibrium condition discussed above if left to do so under its own devices. We show this in the charts below.

EXHIBIT 20: Example Industry Evolution Absent Financial Economy 3500

180 160

Price stability emerges over time and is capable of supporting continued growth in output.

2500

2000

140 120 100 80

1500

60

Price - Arbitrary Units

Supply - Arbitrary Units

3000

1000 40 500

20

0

0 0

5

10

15

20

25

30

35

40

45

Output - LHS

50

55

60

65

70

75

80

85

90

95

Price Level - RHS

Source: Bernstein Analysis & Estimates

FUND MANAGEMENT STRATEGY

BERNSTEIN

23

23 August 2016

EXHIBIT 21: Example Price Evolution - Ex Financial Flows

EXHIBIT 22: Example ROCE Evolution - Ex Financial Flows Example ROCE Evolution - Ex Financial Flows

Example Price Evolution - Ex Financial Flows 180

40%

160

35% 30%

120 25% 100

Equilibrium price level

80

ROCE

Price - Arbitrary Units

140

20% 15%

Equilibrium ROCE level

60 10%

40

5%

20

0%

0 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 Time

Time

Source: Bernstein Analysis & Estimates

Source: Bernstein Analysis & Estimates

As a consequence of the extended time over which the equilibrium reaches "fair value" supernormal profits will be generated should the industry be left to its own devices. These profits clearly have a value and will therefore attract capital into the industry beyond the organic self-generated cash-flows. Thus the financial markets will be able to accelerate the development of a price equilibrium through the provision of additional capital accumulated from the saving of output generated in other parts of the economy. Having made this initial observation, we can then look at how different models of financial market operation can impact the real economy. Obviously this is a huge topic and the analysis we present does not pretend to any great level of sophistication nor is intended to be comprehensive. However one way of categorising the active versus passive elements in the market is by nature of how they respond to information. The notion of "fair value" is critical in this division. It was the importance of establishing that such a "fair value" does indeed exist that motivated the preceding section of analysis. We believe that it is possible to define the "fair value" of any asset based on the mean reverting properties of the price level discovered naturally by any industry operating under its own internal re-investment dynamic. However, given that such a price level exists then, rather than waiting for such a price level to emerge naturally, it is possible to seek to actively precipitate its emergence. By having an understanding of where the "fair value" level ought to be it is possible to seek to profit from this through investing capital (or withdrawing it) in proportion to the extent that the prevailing price level departs from that "fair value level" at any point in time. This requires an investment philosophy that is "active" in so far as it is forward looking and grounds its investment decisions in an effort to understand the dynamics of the real economy. We would say that the "passive" style of investment is paradigmatically opposed to this approach. Rather than looking at the real economy and seeking to understand its future development, passive asset allocation self referentially looks to the financial economy to inform its asset allocations choices. It is necessarily backward looking rather than forward looking and invests based on information gathered from how other financial agents have invested over some historic horizon. Based on this, it does not seek to discover "fair value" itself but simply seeks to allocate more capital to those sectors which appear to be out or underperforming based on the recent past. We model the difference between these two styles as follows. 

Active. Increases (or decreases) the capital invested in an industry beyond that generated organically by the industry incumbents – in proportion to the degree to which the prevailing price level departs from the forward looking "fair price".

FUND MANAGEMENT STRATEGY

BERNSTEIN

24

23 August 2016



Passive. Increases or decreases capital invested in an industry in proportion to the historic price performance of the industry irrespective of concerns regarding fair value.

Under both regimes we have an incremental flow of capital into the real economy; we model that increment as follows (denoting the equilibrium price level as P hat).  

∆𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑡𝑡𝑎𝑎𝑎𝑎𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 ~𝜇𝜇 ∙ �𝑃𝑃𝑡𝑡 − 𝑃𝑃� �

∆𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 ~𝜈𝜈 ∙ (𝑃𝑃𝑡𝑡 − 𝑃𝑃𝑡𝑡−1 )

We can then model the impact of these two regimes on the performance of the real economy. Once this is done it is possible to look at a further refinement of this model which acknowledges that the state of the financial economy is itself in flux and that there is a combination of both active and passive allocation of capital at work with a weighting (say α) between the two giving the total non-industrial capital at work in the industry. ∆𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 = 𝛼𝛼 ∙ ∆𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 + (1 − 𝛼𝛼) ∙ ∆𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 Under the first dynamic, that of active capital management, the following picture emerges. EXHIBIT 23: Example Price Evolution - Inc Financial Flows

EXHIBIT 24: Example ROCE Evolution - Inc Financial Flows

Example Price Evolution - Inc Financial Flows

Example ROCE Evolution - Inc Financial Flows

180

40%

160

Price - Arbitrary Units

140 120 100 80

35% 30% 25%

ROCE

Equilibrium price level much more rapidly discovered by the existence of financial markets

Again, the more rapid discovery of equilibrium

20% 15%

60 10%

40

5%

20 0

0% 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

Time

Source: Bernstein Analysis & Estimates

Time

Source: Bernstein Analysis & Estimates

Under the dynamic of active management, we see the industry discover its natural equilibrium far more rapidly than would otherwise have been the case. The financial markets act to accelerate the performance of the real economy and to bring forward in time the level to which the real economy would revert if left to act under the reinvestment of its own organically generated capital. There is a clear public, social and industrial good served by the financial markets under this regime. Under the second regime, that of passive management, the following picture emerges.

FUND MANAGEMENT STRATEGY

BERNSTEIN

25

23 August 2016

EXHIBIT 25: Example Price Evolution - Inc Financial Flows

EXHIBIT 26: Example ROCE Evolution - Inc Financial Flows

Example Price Evolution - Inc Financial Flows

Example ROCE Evolution - Inc Financial Flows

180

45%

Under the dynamic of passive financial capital allocation, a source of extreme price volatility is introduced

Price - Arbitrary Units

140 120 100 80

40% 35% 30%

ROCE

160

20%

60

15%

40

10%

20

5%

0

Again when looking at financial returns on investment.

25%

0% 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95

Time

Source: Bernstein Analysis & Estimates

Time

Source: Bernstein Analysis & Estimates

Under this regime, a significant source of volatility is introduced. Eventually the real economy is able to find equilibrium but this occurs despite rather than because of the financial economy. (In the above model we have the real economy continuing to invest according to the residual cash flows generated by a fixed dividend pay-out ratio, so in this sense the real economy disregards the capital flows generated by passive asset management). However the time horizon over which this occurs is extended and there is a vast increase in wasted resources – in the form of price volatility – that is engendered by the capital markets. Under this dynamic there is no public, social or economic justification for the capital markets. So what is going on here? Well ultimately it comes down to whether the fundamentals of the real economy or momentum in the financial economy that drive price discovery. Under the first alternative the capital that flows in (or out) of a sector is dependent on an assessment of future needs. It is an intrinsically forward-looking and consequently self-correcting mechanism. Of course the assessment of what those future needs are likely to be is far from static, but that is a different issue. Under the second alternative (that of passive financial markets) capital inflows are not determined by forward looking fair value but by backward looking momentum. Under this dynamic there is no self-correcting price mechanism, indeed the feedback look so established is entirely negative. Capital inflows are greatest when past performance has been highest, which is itself unrelated as to whether that past performance is warranted or not. Hence there is a dilution rather than a concentration of information about the needs of the real economy. We will touch on some comments made by Andrew Haldane on the nature of markets later in this note, but the dynamic we illustrate here highlights exactly some of the points he raised. Fundamentally policy makers need to decide whether markets are a mechanism for price discovery or for punishing dissent? They will never be wholly one or the other, but the role of passive management has tilted the balance of their role in one direction and the implications of that rebalancing need to be understood. The forgoing analysis assumes a perfectly determinate world, which is clearly at odds with empirical experience. Demand is never a constant but is instead an essentially stochastic process. There is significant new information that flows into the market all of which has an ability to change an understanding of where fair value for the real economy lies and the level of the price equilibrium that should result. Expectations around growth rates change, variations in the exchange rate environment and macro level inflationary (and deflationary) pressure can significantly change our understanding of the cost structure of an industry as can innovation and unanticipated technological shifts. All of these result in a situation where significant price volatility is generated. However the volatility engendered by this process, i.e. the process by which new information is factored into an assessment of fair value is of a fundamentally different nature to the volatility created by the structure of how financial markets interact with the real economy. While in practise it may be difficult, indeed impossible, to disentangle the causes of volatility this does not alter the fact that a difference does indeed exist. The volatility attributable to active management trying to

FUND MANAGEMENT STRATEGY

BERNSTEIN

26

23 August 2016

understand the fair price equilibrium required in the efficient function of the real economy has a social good, the volatility associated with passive inflows of capital has no such justification. With these points in mind we would make the following observations about the importance of efficient financial markets to the functioning of the most important part of real economy – namely the primary extractive industries. In the first place that the role of the financial markets in supporting the real economy can be very roughly divided into two stages. 

At the initial introduction of a new technology or market where existing reserves of capital are insufficient to support the full development of an innovation.



Once a market is established to ensure that the level of capitalisation of the industry remains appropriate and in particular that the organic cash flow generated by the industry is either reinvested or returned to investors in the most appropriate (value maximising) manner.

With regard the first of these processes, the following quote provides a perfect illustration of the mechanism at work.

"Certain experiments indicated the probability that ores running (as low as) 2 percent in copper might be worked at a profit. The result was the initiation of the project to mill these ores on a scale which took people's breath away, for it was decided to put in a mill to treat 6000 tons a day and even 12,000 tons a day was talked of. In this connection it is well to bear in mind that the peculiarity of the disseminated deposits (porphyries) as compared with the older mines of copper, was the immediate requirement of large sums of money for the necessary construction of plants and developments of properties before production could begin. A necessary factor in success was a large scale of operations. Undoubtedly the projectors of the first porphyry mines felt somewhat appalled at the risks they were taking in asking for the amount of capital required to launch these enterprises. At this time a mill that would concentrate 1,000 tons a day was considered a pretty large one, but such a mill applied to the low grade disseminated ores would scarcely make any profits at all. A mill of 5,000 or 10,000 tons a day seemed justified. But it cost a great deal of money to build such plants, vastly more money than the projectors were able to furnish. The result was that they had to appeal to bankers…" This was written in 1920 and describes the process of the capitalisation of one of the most important technological innovations in the history of the modern economy, namely the development of the copper porphyry deposits of the USA at the turn of the twentieth century. However for mining, and indeed for other capital intensive sectors of the economy, it is no longer in the requirement to source equity for initial capital investment that the efficient functioning of the capital markets is important. Very few of the major mining companies actually raise significant amounts of cash directly through the equity markets. However it is the market value of the companies that determines the level of indebtedness that they can support and thus their overall spending plans and investment rate. So if anything this role requires a more rather than less efficient functioning of the equity markets. There are a number of observations to make with regard to mining. 

Mining is systemically important. As the origination point (along with energy) for all supply chains in an economy there is a degree of importance attached to this activity that is not common to other industries. Tertiary sector activities no not have this property and are consequently less systemically important. Whether Facebook survives or not is a matter of very little economic significance, the same cannot be said for Rio Tinto. Absent Facebook, aggregate productivity probably rises, absent Rio then the world very quickly moves to a Mad Max mode of existence.



Mining is capital intensive. To the extent that an industry is not capital intensive then by definition capital allocation decisions are relatively unimportant. And if this is the case the role of efficient capital markets is lessened. However mining is the most capital intensive of all modern economic activities, the capital allocation decisions in mining are, accordingly, of huge importance and therefore the industry is highly reliant on an efficient capital market to help provide information on what those decisions ought to be.



Mining is a long lead time activity. This follows on from the fact that mining is a capital intensive activity. In the real economy there is a limit to how quickly it is possible to deploy capital without undue haste actually eroding its value. It is no simple matter to deploy billions of dollars of capital a year. The higher the capital intensity the longer the lead time before capital can be deployed.



"Alpha" generation is commensurately long lived. As a consequence of these factors the time horizon over which value is created in the industry is extremely long, running to decades or more. This is the horizon over which "alpha" is actually generated and this is the horizon over which wealth creation (rather than redistribution) takes place. In order for a market to be efficient in the sense that is required to help guide the capital allocation decisions in the real economy the financial

FUND MANAGEMENT STRATEGY

BERNSTEIN

27

23 August 2016

market must be able to look through horizons that are shorter lived than this. Clearly this is not an easy task but it is nevertheless the task that must be performed. 

Passive flows are more pernicious in long lead time industries. The root cause of the volatility seen in the economic model above (i.e. with passive financial inflows) is the mismatch between project lead times and flows of capital. Capital flows can originate immediately but the real economy cannot generate a growth in output immediately. This is particularly the case for capital intensive sector of the economy like mining. As a consequence, the volatility introduced by "passive" flows of capital will be more pronounced than in those capital light sectors of the economy.



The rise in passive management took place at precisely the point when greater active management was required. It is a coincidence that the rise in a passive management in the financial economy started to really gain traction just when some of the most significant changes to the functioning of the real economy were also taking place. At a time when intelligence about fair value and price discovery was most needed there was a contraction in the aggregate time horizon over which the market was prepared to look (given that the time horizon of passive capital is by definition zero or, indeed, negative). It is probably impossible to quantify the loss in the real economy engendered by this but it is undoubtedly the case that it was significant. (It must be a significant share of the value loss reflected in asset write downs and impairments but this neglects the opportunity cost of the value of misdirected capital)



The increase in volatility in the financial economy erodes the confidence of the real economy to invest. Despite the desire of central bankers to stimulate growth through ever greater interventions it is clear that the world is still in a liquidity trap. We believe that the rise in passive management of the financial economy must take its share of the blame for this. In essence the effect of growth in passive management is to contract the average time horizon over which the market looks. In so doing it increasingly privileges short term earnings over long term value. At a time when the time horizon over which the real economy operates is expanding, such a contraction makes it harder and harder for the capital intensive sectors of the economy to invest properly. If the price signals generated by the financial economy are misleading, then it is virtually impossible to expect the real economy to have much confidence in the information contained in the market.

What this highlights is that the capital markets have a dramatic ability to influence the process of price discovery. Indeed the financial markets can anticipate the process of price discovery. Given that the scale of the capital available in the financial markets dwarfs that of the physical market (for example, last year the LME traded 41 million lots of copper at 25 tonnes each, or nearly 1 billion tonnes of metal versus mined supply of less than 20 million tonnes) then the inflows and outflows of capital into the market will set the price to reflect what the market expectations of fair value are. Under this formalism it is not that the financial markets are opposed to industry fundamentals, but rather that the financial markets will enable price to move more or less instantaneously to changes in what the market in aggregate regards those fundamentals to be. This would explain why commodities can so often depart from what appears to be the equilibrium price level given the current cost structure of the industry; for example why the iron ore price fell as dramatically as it did in 2015 despite the apparent cost curve support of all the high cost Chinese mining. The price moved to reflect the new market expectations of a slower growth world (albeit perhaps too slow) and did not need to wait for the self-equilibrating action of industry fundamentals to discover this new price level. So this creates one of the most powerful applications of this approach to commodity pricing, in that it enables us to understand what are the fundamental expectations embedded in any commodity price level. Most of the information in the equilibrium price level (shown again below) is readily observable, such as the cost of production and the capital intensity and so forth. 𝑃𝑃𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 = 𝛼𝛼 +

𝐾𝐾 ∙ (𝑛𝑛 + 1) ∙ �(1 + 𝑔𝑔)𝜏𝜏 ∙ (𝑔𝑔 + 𝑑𝑑) − 𝑑𝑑 ∙ [(1 − 𝑇𝑇) ∙ 𝐷𝐷 + 𝑇𝑇]� 𝑛𝑛 ∙ (1 − 𝑇𝑇) ∙ (1 − 𝐷𝐷)

The one variable that is not so easily observable is the market expectations for growth. However this then provides an interesting test for this view of commodity prices, which is to look at the growth rate discounted by the commodity price environment over time versus the actual growth rate that pertained. We show the result of this for the copper price back monthly over the last 30 years.

FUND MANAGEMENT STRATEGY

BERNSTEIN

28

Jan-85 Nov-85 Sep-86 Jul-87 May-88 Mar-89 Jan-90 Nov-90 Sep-91 Jul-92 May-93 Mar-94 Jan-95 Nov-95 Sep-96 Jul-97 May-98 Mar-99 Jan-00 Nov-00 Sep-01 Jul-02 May-03 Mar-04 Jan-05 Nov-05 Sep-06 Jul-07 May-08 Mar-09 Jan-10 Nov-10 Sep-11 Jul-12 May-13 Mar-14 Jan-15 Nov-15

Jan-85 Oct-85 Jul-86 Apr-87 Jan-88 Oct-88 Jul-89 Apr-90 Jan-91 Oct-91 Jul-92 Apr-93 Jan-94 Oct-94 Jul-95 Apr-96 Jan-97 Oct-97 Jul-98 Apr-99 Jan-00 Oct-00 Jul-01 Apr-02 Jan-03 Oct-03 Jul-04 Apr-05 Jan-06 Oct-06 Jul-07 Apr-08 Jan-09 Oct-09 Jul-10 Apr-11 Jan-12 Oct-12 Jul-13 Apr-14 Jan-15 Oct-15

23 August 2016

EXHIBIT 27: Growth Rate Discounted in Commodity Price 18%

16%

14%

12%

10%

8%

6%

4%

2%

0%

-2%

Source: Bernstein Analysis & Estimates

EXHIBIT 28: Actual vs Implied Growth Rates

20%

15%

10%

5%

0%

-5%

-10%

Growth Rate Implied by Copper Price

FUND MANAGEMENT STRATEGY

Actual Growth Rate

Source: Bernstein Analysis & Estimates

BERNSTEIN

29

23 August 2016

EXHIBIT 29: Relationship Between Implied and Actual Growth Rates 16.0%

14.0%

R² = 25%

Implied Global Growth

12.0%

10.0%

8.0%

6.0%

4.0%

2.0%

-10.0%

-5.0%

0.0% 0.0%

5.0%

10.0%

15.0%

20.0%

Actual Global Growth

Source: Bernstein Analysis & Estimates

The t-statistic on the relationship is 3.2 and the P-value 0.3%, which indicates the strength of the statistical relationship at work. What this confirms is that it was the radical change in growth expectations over the course of 2015 that resulted in the collapse in commodity prices despite the apparent cost curve support that many enjoyed. On this basis the market price of copper currently seems to discount a long run growth expectation of 0.6% versus the 30 year average of 6.1%.

FUND MANAGEMENT STRATEGY

BERNSTEIN

30

23 August 2016

CORRELATION, LIQUIDITY, PENSIONS AND ESG: OTHER SYSTEMIC ISSUES WITH PASSIVE MANAGEMENT The first sections of this note focussed on the role of active and passive management in capital allocation. We now turn to other consequences of the rise of passive for the financial system.

THE NEED TO HAVE RISING ASSET MARKETS TO FUND THE PENSION SYSTEM The role that equity markets have in nurturing growth might be seen as a good in itself. But there is another more specific social need when it comes to asset markets. Society needs to have rising asset markets in order to fund the pension system. For DB schemes this is less relevant as they can immunise their portfolios by matching the cash flows of their assets and liabilities, but for DC schemes the assumption is that rising asset markets will provide the pay-off needed. Until now the bulk of this has rested on a long-only allocation to fixed income and equity markets. Whatever one's views are of macroeconomic growth it seems likely that we are overdue a period of low returns given low bond yields and relatively high Shiller PEs which point to sub-par returns both for equities and bonds. Exhibit 30 shows the level of the US 10 year bond yield and 10 year forward returns from US Government bonds. The conclusion is that over the last 216 years the current level of the yield has been a good predictor of 10 year forward annual returns. So with the US 10 year bond currently yielding less than 2%, we should expect less than 2% pa return on US sovereign bond portfolios. What about equities? The best predictor of 10 year forward equity returns is the Shiller PE (price divided by 10 year average inflation-adjusted earnings), Exhibit 31. Its current level tells us that our best forecast of annual total returns (including dividends) for US equities is 6% pa. How does this compare to what pension schemes need to make? The average expected return on plan assets by US corporate pension schemes is 7% pa14. There is no 60:40 combination of sub 2% and 6% that yields 7%. Yes, schemes can invest outside the US and can also to some extent move down the quality spectrum in credit assets but there is still a problem here. Ultimately, in a world where returns are low and non-financial corporates have de-levered it might make sense for asset owners to take on more leverage themselves, but that is outside the scope of this note. This is, ultimately, a major challenge but also the great opportunity for asset managers. They will be relied upon to manufacture the return streams to meet these liabilities.

14

Please see In Defence of Active Management

FUND MANAGEMENT STRATEGY

BERNSTEIN

31

23 August 2016

EXHIBIT 30: USA 10 year bond constant maturity yield and 10 year forward bond total return

EXHIBIT 31: 10 year forward total return and US Shiller PE

18

25

16

5

20

Total Return (USD)

12 10 8 6 4

15

15

10

20 25

5

30

0

35

Ratio (inverted)

10

14

40 -5

2

45

-10

31/01/1800 31/01/1813 31/01/1826 31/01/1839 31/01/1852 31/01/1865 31/01/1878 31/01/1891 31/01/1904 31/01/1917 31/01/1930 31/01/1943 31/01/1956 31/01/1969 31/01/1982 31/01/1995 31/01/2008

0

50

31/01/1881 30/11/1891 30/09/1902 31/07/1913 31/05/1924 31/03/1935 31/01/1946 30/11/1956 30/09/1967 31/07/1978 31/05/1989 31/03/2000 31/01/2011

%

0

USA 10-year bond constant maturity yield

10-year forward total return (annualized)

10-year forward bond total return (annualized)

US Shiller PE (RHS, Inverted)

Source: Global Financial Data, Bernstein analysis

Shiller PE defined as price divided by 10 year average inflation-adjusted earnings. Source: Robert Shiller's database, Global Financial Data, Bernstein analysis: Global Financial Data, Bernstein analysis

The question of interest here is that if a greater share of the equity market becomes passive, can this have a detrimental impact on the performance of the overall equity market itself? There are two routes that this could take, the first is dampening of the overall growth rate of the economy due to a decline in the efficacy of capital allocation as discussed in the previous section. The other is whether the equity market itself achieves a lower level of return given any level of economic growth because it more poorly represents faster growing parts of the economy.

ROLE OF ACTIVE AS A SOURCE OF LIQUIDITY AND HOW PASSIVE MAGNIFIES CORRELATION SHOCKS The point at which any growth in passive could possibly cause a detrimental effect to capital allocation in the economy could be a long way off. Moreover, the effects of any such change may not become apparent for many years. Thus policymakers could perhaps be excused for not having such items high on their to-do lists. But a more tangible impact of a larger passive allocation on the functioning of capital markets may become more apparent at a much earlier stage. That is in the other major role that active management plays which is in liquidity provision to investors. This forms a key part of the most comprehensive study of the impact of passive investing on the functioning of the market which is the recent paper by Bolla, Kohler and Wittig (2016)15. They provide evidence that the rise of passive investing has a measurable impact on financial stability. Specifically they analyse the impact of passive investing on six metrics of risk commonality: 1) Cross-sectional dispersion of changes in trading volume 2) Average pairwise correlation of changes in trading volume 3) Average pairwise correlation of returns 4) Average absolute difference of return betas

15

Bolla, Kohler and Wittig (2016): Index-Linked Investing—A Curse for the Stability of Financial Markets around the Globe? Journal of Portfolio Management Spring 2016, Vol. 42, No. 3: pp. 26–43

FUND MANAGEMENT STRATEGY

BERNSTEIN

32

23 August 2016

5) Average pairwise correlation of Amihud liquidity16 6) Average absolute variation in return betas

They study the relationship between these variables and the share of equity assets managed passively in the US, Eurozone, UK and Switzerland from 2001-2014. Within each market they find a significant relationship over time between their risk commonality measures and the passive share of the market. They also show that the same result holds in comparing risk commonality across markets. The conclusion is that returns and liquidity have greater commonality when the share of assets managed passively is higher. They point out that this increases the probability of adverse tail events and also hinders the ability to achieve diversification in portfolios.

OK BUT PASSIVE SHARE OF AUM IS NOT GOING TO 100% One response to all this is that one might say that this is all very well but passive management is never going to be 100% of the market, so how much does this matter? However, these studies suggest that there can be an economically meaningful impact at penetration levels much lower than that. What would happen if this share rose to 60% of AUM or 80%? We can use the relationship between correlation and liquidity with passive from recent academic work to extrapolate the impact of a further increase in passive asset share. Over the last 5 years the proportion of assets run passively has increased by a third or 10 percentage points globally from 25% to 34%. If we assume the same rate of growth of passive over the next 5 years what would the consequences be? We can use the results from the various academic studies that have been conducted to postulate what the impact on correlation of returns and liquidity would be if the recent trend of the growth in passive increases further. Using the data from the Bolla, Kohler and Wittig (2016) paper a further one third increase in passive share would increase the level of average correlation between stocks by 16pp in the Eurozone and 14pp in the US. The level of correlation in Europe in particular is already elevated so this would push it to a level that we have not previously seen before even in the depths of the financial crisis, Exhibit 32. In the US the correlation would be pushed up to a level that was only exceeded for short periods during the crisis, Exhibit 33. But this is the increase in correlation that would result from the increase in passive AUM alone, were an exogenous shock to act on the market then it could push correlations still higher. Likewise the cross-sectional dispersion of changes in trading volume would be expected to decline by 12pp in the Eurozone and by 6% in the US17.

16

Amihud, Y (2002): Illiquidity and stock returns: cross-section and time-series effects Journal of Financial Markets Volume 5, Issue 1, January 2002, Pages 31–56 17 We derive these estimates by using the coefficients derived by Bolla, Kohler and Wittig (2016) for the dependence on risk commonality factor on ETF AUM.

FUND MANAGEMENT STRATEGY

BERNSTEIN

33

23 August 2016

EXHIBIT 32: European stock correlation and impact of increasing passive

EXHIBIT 33: US stock correlation and impact of increasing passive

0.6

0.6

0.5

0.5

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1

0

0

03/01/2000 03/01/2001 03/01/2002 03/01/2003 03/01/2004 03/01/2005 03/01/2006 03/01/2007 03/01/2008 03/01/2009 03/01/2010 03/01/2011 03/01/2012 03/01/2013 03/01/2014 03/01/2015 03/01/2016

0.7

03/01/2000 03/01/2001 03/01/2002 03/01/2003 03/01/2004 03/01/2005 03/01/2006 03/01/2007 03/01/2008 03/01/2009 03/01/2010 03/01/2011 03/01/2012 03/01/2013 03/01/2014 03/01/2015 03/01/2016

0.7

EU correlation

US correlation

Estimated correlation for a 1/3 increase in passive asset share

Estimated correlation for a 1/3 increase in passive asset share

Figure shows the average pairwise correlation of stocks based on a rolling six month window. The point shows an estimate of the impact on stock correlation of a 1/3 increase in the proportion of assets run passively. This uses the coefficient for the dependence of stock correlation on passive asset share derived in Bolla et al (2016). Source: Bernstein Analysis, Bolla et al (2016), Journal of Portfolio Management.

Figure shows the average pairwise correlation of stocks based on a rolling six month window. The point shows an estimate of the impact on stock correlation of a 1/3 increase in the proportion of assets run passively. This uses the coefficient for the dependence of stock correlation on passive asset share derived in Bolla et al (2016). Source: Bernstein Analysis, Bolla et al (2016), Journal of Portfolio Management.

If an increase in the background level of correlation results from the rise of passive investment then the Wurgler (2000) research can be used to gauge the impact on capital allocation. This suggests that a 1 standard deviation increase in the degree to which stocks co-move is associated with half a standard deviation decrease in the elasticity of capital allocation (ie a decrease in the responsiveness of capital allocation to the growth potential of industries). Note here we assume that the impact of passive on correlation, liquidity and capital allocation is linear as that is the conservative thing to do. Those that are more fearful of the rise of passive often claim that at high levels of passive penetration the impact could become non-linear (ie that the impact would be greater than that stated here), but we do not address that here as, to our knowledge, no one has been able to derive a theoretical threshold at which point such effects may occur.

WHO SHOULD HAVE THE FIDUCIARY RESPONSIBILITY FOR FACTOR ALLOCATION? A significant body of policy initiatives is forming around the topic of fiduciary responsibility and the model of how retail investors pay for financial advice. The change in this regard currently attracting most discussion is the US Department of Labor proposed 18

changes to the fiduciary rules for retirement accounts due to take effect in April 2017 . This proposal would require all advisors to act as fiduciaries when making recommendations and/or giving advice on 401(k) plans or Individual Retirement Accounts. There is a view that this could hasten the broader movement from commission to fee-based payments. In that sense it has some similarities to the Retail Distribution Review (RDR) in the UK which came into effect in 2013 which required retail clients to pay fees rather than pay commissions. There have also been similar moves in the Australian market with the Future of Financial Advice reforms19 from 2012 also imposing a best interest duty in the provision of advice to retail clients and a shift to more fee rather than commission-based payment.

18 19

https://www.dol.gov/ebsa/newsroom/fsconflictsofinterest.html http://asic.gov.au/regulatory-resources/financial-services/future-of-financial-advice-reforms/

FUND MANAGEMENT STRATEGY

BERNSTEIN

34

23 August 2016

Much debate around these reforms has centred on whether it creates an "advice gap" for less well off investors who do not wish to pay an explicit fee for the advice. We will return to that particular topic in future research, but for our purposes here the more germane issue is what is the nature of fiduciary responsibility in a world where the passive-active distinction has become blurred. In the rush to passive some active decisions are being made but no longer explicitly recognised as such. This is the Achilles heel of smart beta: who should have fiduciary responsibility for factor allocation? If a pension fund, or indeed a retail investor, divests from an active fund that was giving a bundle of returns and instead buys a "passive" smart beta ETF several things occur. They probably lower the headline fee that they pay but they also could in some cases see a shift in the composition of their returns from a blend of factors and a mix of idiosyncratic and systematic components to a bullet exposure on one or a small number of factors. That is a highly active factor allocation decision but we worry that it is not always explicitly recognised as such. We suggest that there could be an opportunity for asset managers here. Yes the most sophisticated sovereign wealth funds and pension funds can make this active factor allocation themselves, but for others this is probably an area where a combination of the asset allocation and solutions groups within asset managers can and should take market share.

ESG We think that the growth in ESG or SRI investments is a key part of the defence of the role of active and in particular of outlining its social role. In the initial stages of ESG investment there was a flurry of research conducted on whether such investments outperform or lower risk. We think that making that the primary thrust of an ESG approach is misguided. Firstly, the evidence for outperformance is weak at best (there is possibly a case for lowering risk in some cases) but more importantly it does not usually reflect the reason for setting the ESG mandate. There are several reasons why such mandates may be set: 

A desire to improve the environmental or social outcomes from investing to help in the creation of a better world in some way



A belief that in the long run either through impact on the environment or through regulation there will in fact be an eventual underperformance of stocks that score poorly on these metrics



A wish to avoid negative publicity from holding assets that are deemed morally questionable

It is possible that all three reasons may be cited for a shift into ESG. The point that is germane to our subject here is that if someone (be it an asset owner or a government) sets a mandate for managing assets that includes an ESG target for the first reason, ie to aid in the creation of a better world, then they are implicitly making a strong case for active investing. The assumption has to be that the process of capital allocation can be influenced by taking the active decision to divest from or to underweight certain companies or sectors. Here the distinction between active and passive can be subtle. It is perfectly possible to make an ESG investment where the implementation is entirely passive in the sense that assets are directed into an index that follows simple systematic rules with no discretion on the part of the manager. However there are still two active and discretionary parts to the decision: 1) The decision by the asset owner to allocate into such an index (ie an active asset allocation or factor allocation decision) and 2) The choice on the part of the index constructor how the rules for the ESG index in question should be set. The implementation process could then be passive or active. Although the growth of ESG mandates is much discussed we find it odd that it is not generally recognised as a defence of active management in its own right. This could be a significant support for the active industry in coming years. The key point for asset owners is probably how such mandates should be set, and for asset managers it is what they need to do to win such mandates, a subject that we will return to in upcoming research.

THE NEED FOR PATIENCE We think that the need for more patience in investment is another area that should attract the interests of policymakers as well. The reason why this might be an area for policymaking is that for any individual market participant (eg a consultant, fiduciary advisor, asset owner) they may be acting rationally by making an individual decision to buy or sell a fund in the same short term way as everyone else, but for the system overall this is likely to be suboptimal. There are two distinct issues here:

FUND MANAGEMENT STRATEGY

BERNSTEIN

35

23 August 2016



The churn costs that create a persistent wedge between the average return of active funds and the average return achieved by investors in those funds.



If active funds lose market share every time they underperform that will encourage a shortening of the active investment horizons which in turn will also magnify the misallocation of capital. Therefore is the rise of passive making the active investing that remains less able to fulfil its capital allocation role?

There are two reasons why the achieved return on the part of investors can lag the average return of funds: poor market timing and poor fund selection. For example if end investors on average divested from equities when they fall in such a way that they miss out from rebounds in the market then their performance would lag the market. Likewise if investors divested from managers who underperform in such a way that they missed out from any subsequent outperformance of that manager then their performance would lag the average performance of managers. The first element of this churn cost is common across investors in both passive and active funds while the latter is an additional issue for investors in active. If investors in active funds buy and sell those funds at inopportune times then their perceived experience of active investment will lag behind its potential. What can we say about the scale of this? Evidence suggests that the fixation on 3 year track records has a major effect, with average holding period in mutual funds is a little over three years both within equities and fixed income (Exhibit 34). Investors in cross-asset funds tend to be more patient and wait closer to 4½ years (presumably because there is less of an obvious benchmark for them to be compared against).

EXHIBIT 34: Average mutual fund holding periods 5 4.3

4.5 4 3.3

3.5

3.2

3 2.5 2 1.5 1 0.5 0 Equity Funds

Bond Funds

Balanced Funds

Source: Dalbar 2012

This is particularly detrimental as this time frame for fund selection appears to be one over which on average there are reversals in terms of fund outperformance. Almost by definition managers who are hired tend to outperform over the three years prior to being hired and fired managers underperform over the three prior years, Exhibit 35.

FUND MANAGEMENT STRATEGY

BERNSTEIN

36

23 August 2016

EXHIBIT 35: Fired managers tend to outperform hired managers in the three years after the hiring/firing decision

Data: 8,775 hiring decisions by 3,417 plan sponsors delegating $627 billion in assets; 869 firing decision by 482 plan sponsors withdrawing $105 billion in assets. Analysis covers the period 1996 through 2003. Source: "The Selection and Termination of Investment Management Firms by Plan Sponsors," Amit Goyal, Sunil Wahal (The Journal of Finance, Volume 63, Issue 4, printed August 2008). 20 What does this mean for investment returns? Morningstar estimates that over the 10 years ending December 2015 the average investor in International equity funds achieved 3.91% pa return while the average fund achieved returns of 2.67% pa. The gap between these numbers being the "churn cost", it the combined effect of poor equity market timing and poor timing in selection of funds on the part of investors. This implies an annual churn cost of 1.24% pa. The gap is slightly smaller for the US funds at 0.74% pa.

This is a system-wide problem. It might make sense for individuals responsible for fund allocation to churn funds in this way but for the end asset owner and for society at large it is not optimal. Maybe investors need to be "saved from themselves" and somehow behaviour should be encouraged that slows down fund churning. We suggest that this would be better both for asset owners and for asset managers (although less for middlemen): asset owners would be saved from at least some of the considerable churn cost discussed here and asset managers would be freer to pursue investments that better accord with their views as opposed to constantly looking over their shoulder. The costs churning funds that investors impose on themselves is only one part of a time-horizon problem associated with active investment. There is a debate about whether financial markets have become too short-term which has attracted the attention of policymakers. The rise of passive contributes to this as it makes it very obvious when active underperforms even for short periods. If active loses assets every time it underperforms there will be a ratcheting effect of outflows from active funds that will contribute to a myopic over-focus on short-term performance by fund managers and also by those who are responsible for allocating to such funds. This in turn has to impact the investment decisions of the fund manager and could stay their hand from implementing investments that they may otherwise make. If we believe that active management does indeed contribute to the process of capital allocation then this shrinking of the measurement horizon for managers will impair that process and make active management less able to fulfil its capital allocation role. Andrew Haldane in his speech "Patience and Finance"21 reminds us that liberalised markets can have two possible equilibria, one in which patient long term investing dominates and in which prices mean-revert over time to reflect fundamentals or the other in which impatience takes over and with it momentum as a strategy dominates, then investors who do not follow the herd are punished by having their funds removed. These two possibilities have large implications for the type of investment strategy that can succeed over time (eg can mean-reversion work?) and also for the role that investment activity can have in capital allocation. 20

See MorningstarAdvisor, Mind the Gap 2016, Russel Kinnel https://corporate1.morningstar.com/ResearchArticle.aspx?documentId=756760 21 www.bankofengland.co.uk

FUND MANAGEMENT STRATEGY

BERNSTEIN

37

23 August 2016

On a more formal basis, The Kay Review of UK equity markets and long-term decision making (2012) demonstrated that shorttermism is an issue in the asset management industry and worried that it means that the ultimate asset owners were losing out. The review suggested that the relationship between ultimate savers and institutional investors is subject to principle-agent problems that can give rise to a myopic behaviour of the latter. For example, if the ultimate holders of investments use shortterm performance indicators to monitor asset managers or formulate short-term targets in investment mandates, then asset managers have an incentive to favour short-term profits over long-term profitability. The Kay Review recommended discouraging the use of measures and models that rely on short-term volatility of returns and deviations from indexes when deciding on the remuneration of asset managers22. The conclusion from all this is that the evolution of a system whereby any underperformance relative to a very visible passive alternative leads to a redemption from active funds can shorten investment horizons and in itself harm the ability of active management to fulfil a capital allocation role.

CONCLUSION: IS THERE A LIMIT TO THE SIZE OF PASSIVE OR A NATURAL EQUILIBRIUM BETWEEN ACTIVE AND PASSIVE? We are often asked if we are close to reaching a limit to the size of passive investment, or whether there will be some natural mean reversion to eventually favour active. Proponents of such a view often seem to want to extrapolate this to suggest that one day we will wake up and discover that we are in some kind of active nirvana where the size of passive investments has made the market so inefficient that opportunities for adding value through active investing abound. We are very sceptical of such claims and think they are wishful thinking on the part of active managers. There are several reasons why we think we are nowhere near a limit for the share of the market that is passive: 

The growth in market share of passive equities has been monotonic for the last 10 years, Exhibit 36. This growth of passive has demonstrated no link to whether there are net inflows or outflows to equities nor even to changes in the average pairwise correlation of stocks. The correlation of stocks matters in this case as it is a proxy for the addressable opportunity set for active management as defined by the co-called fundamental law of active management due to Grinold23 which states that 𝐼𝐼𝐼𝐼 = 𝐼𝐼𝐼𝐼. √𝑁𝑁, ie that the Information Ratio (IR) of a strategy is determined by the number of independent investment ideas, N, and the information coefficient (IC) or skill of the manager in identifying them. If investors were rationale they might prefer to invest in active management when correlation was low or falling and prefer passive if correlation was high or rising, yet aggregate passive share does not seem to reflect this.

22

Kay review 2012 www.gov.uk. Also relevant to the discussion of an increased focus on fiduciary standards, the review also called for study of the concept of a fiduciary in investments and for fiduciary standards to be applied to "all relationships in the investment chain which involve discretion over the investments of others, or advice on investment decisions. These obligations should be independent of the classification of the client". 23 Grinold, R. (1989) The Fundamental Law of Active Management The Journal of Portfolio Management, vol. 15, no. 3 (Spring): 30–38.

FUND MANAGEMENT STRATEGY

BERNSTEIN

38

23 August 2016

EXHIBIT 36: The growth of passive has been monotonic 20 18 16 14 12 10 8 6 4 2

Jan-16

Jan-15

Jan-14

Jan-13

Jan-12

Jan-11

Jan-10

Jan-09

Jan-08

Jan-07

Jan-06

Jan-05

Jan-04

Jan-03

Jan-02

Jan-01

Jan-00

0

Source: EPFR Global, Bernstein analysis



The share of assets run passively is 35% globally but this masks some large regional differences. In the US passive (passive ETFs and passive funds) account for 40% of fund AUM, and in Switzerland 50%24, but in Europe and global EM it is 30%. So presumably the passive penetration for non US markets could rise to US levels without massive short-term ill effects.



Likewise, passive penetration of large stocks is much larger than for smaller stocks. In the US 40% of funds invested in the S&P 500 are passive, but for smaller stocks in the Russell 2000 only 16% are passive. Presumably passive share could increase for smaller stocks (with possible caveats for greater information asymmetry for smaller companies).



Passive accounts for 35% of all equity fund assets but funds only account for a portion of the equity market as equities are also held by other participants: direct holdings by investors, corporates, governments and sovereign wealth funds etc. So as a proportion of total equity market cap passive is just 8.7% of global equities which, while still large, suddenly seems less significant.



All the statistics on the size of passive quoted in this section relate only to passive in the traditional cap-weighted sense. However we think that smart beta should count as passive too. Thus whatever one previously thought might have been an upper limit on the size of passive has become larger as smart beta includes assets with a correlation less than one with the rest of passive.



If we were getting close to some natural limit for the size of passive it seems likely that this passive share would have shown some let-up or alteration.

24

Mikkelsen (2016): European Fund Expenses Are Decreasing in Percentage, But investors pay more in nominal values, MorningStar 2016 http://media.morningstar.com/uk%5CMEDIA%5CResearch_Paper%5C2016_Morningstar_European_Cost_Study_1708 2016.pdf

FUND MANAGEMENT STRATEGY

BERNSTEIN

39

23 August 2016

The question ultimately is how large can passive get? Can it become too large? In fact we think that this question is not specified well enough to actually answer. The more pertinent question is: too large for what? For market efficiency? For Capital allocation? For liquidity or correlation? Phrased in these ways it might be at least possible to measure the impact from the growth of passive investment and establish whether there is demand for a counteracting growth of active management. Specifically the questions then become: 

Has the market become inefficient in some sense such that active opportunities become good enough for large returns, the "active nirvana" result?



Has the size of active become so small that Capital allocation in the economy breaks down in some way?



Has passive management become so large that is has a detrimental impact on liquidity or correlation?

With the first two questions one would have to ask "How would one know if such a point had been reached?". It could be far from obvious, at least until the passage of time was great enough that one could look back and point a finger at some past moment. While the final question might be more possible to answer contemporaneously, it is far from obvious what any individual institution could actually do about it.

… OR DOES THIS NEED POLICYMAKERS TO SIT UP AND CARE? The first two questions above require the passage of time to make their impact felt and the latter two are questions for the system overall rather than something that can be addressed by a single investor. We are often asked whether we reach a point at which the market becomes so inefficient that the opportunities for active management become unstoppable in forcing an active recovery. So do we ever reach this "active nirvana?" maybe not. Phenomena that only become evident long after the event do not easily lend themselves to quick mean-reversion and individual participants would seem to have little power to bring to bear on the point. Moreover, there might not be any natural meanreversion because the commercial imperative of passive and active asset management is very different when it comes to scale. These businesses have a different natural size. Passive management requires scale in order to cut the fees charged to the levels that we see today and this pressure will always be there. Active management, by contrast, will always have a capacity constraint. The size of that constraint will be different for a concentrated 15 stock equity fund versus a multiasset index fund, but both of those strategies have constraints nonetheless. Thus the industry might not have a self-correcting equilibrium process between active and passive given these different forces at work in terms of natural scale. If there is no natural mean-reverting mechanism does it need regulators and governments to get involved? (or, as one US client recently remarked when your author suggested this in a meeting, is that just a very "European" response?). When there is a possibility of market failure occurring, that is the point where there may be at least a prima facie case for interest from policymakers. Moreover, as we suggested at the beginning of this note, given the role of policymakers is to consider the broader utilitarian role of an activity and the difficulty of any individual market participant in changing the status quo there could be a role for policy in making sure that active management does not shrink in a way that would be inimical to society at large. We do not for a moment suggest that policymakers should consider limiting passive in any way. That would be detrimental to many asset owners, heavy-handed and anyway probably impossible to do. But instead they may wish to consider the broader benefits of a functioning active asset management industry to society as a whole so that when policy initiatives are undertaken they do not explicitly undermine active management. Thus considering a disinvestment from active management by public pension funds as the UK Government toyed with recently25 or forcing a focus on headline fee above all other considerations may not be in the best interests of the asset owners and beneficiaries in question but also for society as a whole. In fact the massive focus on headline fee rather than desirability of long run return and risk outcome is one case in point. Likewise the ESMA proposal for clamping down on closet indexing might be "fighting the wrong war"26. The growth of passive over the last 10 years has been a huge benefit to asset owners in enabling them to cut their costs. Likewise our view that the next big growth in passive will come from regarding all simple factor products as passive too27 will 25

UK Department for Communities and Local Government: Local Government Pension Scheme: Opportunities for collaboration, cost savings and efficiencies, consultation May 2014 26 Please see Fund Management Strategy: Closet Indexing - Should asset managers care? 27 Please see In Defence of Active Management

FUND MANAGEMENT STRATEGY

BERNSTEIN

40

23 August 2016

take costs down further. This passivisation of factor investing will in turn perform an even more important function in making it clearer for asset owners which parts of active management are most beneficial to them, ie which products are really worth paying for28. But while we recognise that the active vs passive investing debate is nowhere near any policymakers to-do list, we suggest that when decisions on fiduciary responsibility, stewardship, market structure and management of public pension assets are taken that at least the broader benefits to society from active allocation of capital are taken into account. Bagehot may have been writing 140 years ago but his views on the importance of efficient capital allocation are just as valid today as they were then.

A READING LIST We like to end our larger notes with a reading list for readers who wish to delve deeper into the topics discussed.

Equity markets, capital allocation and importance to society. We started our discussion of the importance of capital allocation in society with Bagehot (1873) Lombard Street: A description of the money market. On the big questions of what drives growth in economies and the interaction of that with the functioning of society there are many publications, but two great works on this topic are Landes (1998): The Wealth and Poverty of Nations and Clark (2007): A Farewell to Alms. Both cover economic history from before the industrial revolution to the present day. Landes takes as an implicit focus the question of why some countries are rich and others poor and in Clark there is a long discussion of the causes of the industrial revolution and its impact on per capita income backed up by a wealth of impressively long data sets. Although Solow (2007): The Survival of the Richest? in his review of Clark for the New York review of books which is worth reading in its own right - takes issue with the lack of hard evidence for some of Clark's claims pinning the cause of the industrial revolution on the relatively higher birth rates amongst the richest parts of British society, the amassed data and elements of the argument are highly enriching in any case. Both Landes and Clark at the end of their works turn to the question of inequality which has now received renewed focus with Piketty (2014): Capital in the twenty-first century with its own mass of data that has propelled the topic to perhaps the top of the (long run) economic agenda. We have not discussed the topic of inequality in this note, but a topic for future research would be whether passive management, by potentially aiding an entrenchment of existing firms, has a relative benefit to returns on capital or returns on labour. There is also a literature on the role of financial markets in the allocation of capital. See for example Wurgler (1999): Financial Markets and the Allocation of Capital that presents a very interesting way of showing the link between the efficacy of financial markets and capital allocation through the elasticity of capital allocation linked to how well a market functions. For example he finds that an increase in synchronicity (akin to correlation) leads to a decrease in the responsiveness of capital allocation to changes in profit growth. Levine and Zervos (1998): Stock Markets, Banks, and Economic Growth: Do well-functioning stock markets and banks promote long-run economic growth? shows that both stock market liquidity and banking developments are important in predicting growth and productivity and that thus one ideally wants both forces at work in an economy. Also, Rajan and Zingales (1996) show that industrial sectors that are more dependent on financial markets for raising external finance grow faster in countries that have more developed financial markets.

Impact of passive on the functioning of markets There is now a growing literature on active versus passive allocation and its impact on the market. Bolla, Kohler and Wittig (2016) is the most recent and comprehensive paper on this but also of interest is Wermers and Yao (2010): Active vs passive Investing and the Efficiency of Individual Stock Prices, which finds that stocks with high levels of passive ownership exhibit more long term price anomalies and larger price reversals. Lieppold, Su and Ziegler (2015); Do Index Futures and ETFs affect Stock Return Correlations, suggest that demand shocks to ETFs have a larger impact on price comovement than futures. Sullivan and Xiong (2012): How Indes Trading Increases Market Vulnerability, details how the increase in passive investing appears to lead to an increase in average stock correlations and a convergence of betas across the market. Brogaard, Ringgenberg and Sovich (2016): The Real Impact of Passive Investing in 28

Please see What is worth paying for in an asset manager?

FUND MANAGEMENT STRATEGY

BERNSTEIN

41

23 August 2016

Financial Markets, looked at distortions to commodity markets from an increase in passive investing. They analysed the impact on firms that are users of commodities of the increase in assets that passively track commodity indices. They find that the decreased information value from commodity process that results creates a real cost for such firms. Vayanos and Woolley (2016): The Curse of Benchmarks, that discusses the problem with over-obsession on performance relative to a benchmark. Also of interest is Wurgler (2010): On the Economic Consequences of Index-Linked Investing, that details the proliferation of indices, their impact on price co-movement and the impact that can have via use of the CAPM on real world capital budgeting decisions. We very much like the discussion on the Philosophical Economics blog that often takes the opposite side to these arguments. Notably making the case in Philosophical Economics (2016a) that index Investing could make markets and economies more efficient with its claim (which we would dispute) that any incremental move into passive improves capital allocation by removing the asset manager who is at the margin less skilled. There is also an interesting piece, Philosophical Economics (2016b): The value of active management: a journey into indexville which tries to assess what it would cost to replace the capital allocation function performed by active management with a pool of analysts (as opposed to, say, the management of a collective in a Marxist system as we discussed in this note).

Public Policy, churn costs and active management: We referred to various policy initiatives in this note. The EU has set out the economic analysis that underpinned the Capital Markets Union initiative in its EU commission staff working document (2015). The UK Department for Communities and Local Government commissioned a report into the structure of the local pension schemes in the UK that led to a suggestion (later dropped) to divest from active management, see Department for Communities and Local Government (2013): LGPS Structure Analysis. The current proposed changes in rules for fiduciary responsibility for retirement accounts in the US is available at US Department of Labor (2016): Fact Sheet: Department of Labor Proposes Rule to Address Conflicts of Interest in Retirement Advice. Kay (2012): The Kay review of UK equity Markets and Long-Term Decisions Making worried that ultimate asset owners were losing out because of short term decisions in fund management allocation decisions. For the bigger picture aspect of policy there is also Haldane's excellent speech on the need for Patience – Haldane (2010). For a fantastic and very wide-ranging view on how investment management should adapt see Blanqué (2014): Essays in Positive Fund Management. This covers a much broader array of issues than the ones in this note but of particular relevance to our discussion here it includes topics such as the diversifying factors vs asset classes; whether all investments are in fact active and organisational issues that arise for asset managers from the growth of factor investments. The churn costs that can drive a wedge between the average return of managers and the average return experienced by investors have been studied by a number of authors. MorningStar have the latest publication on this in Kinnel (2016): Mind the Gap 2016 and there is also Quantitative Analysis of Investor Behavior from DALBAR. Amit Goyal, Sunil Wahal (2008) separately show the average performance of mangers before and after hiring and termination decisions.

Bibliography Aguilar (2015): U.S. Equity Market Structure: Making Our Markets Work Better for Investors, SEC Public Statement. May 11, 2015. Amihud, Y (2002): Illiquidity and stock returns: cross-section and time-series effects Journal of Financial Markets Volume 5, Issue 1, January 2002, Pages 31–56 Bagehot (1873) Lombard Street: A description of the money market, available at https://books.google.co.uk/books?id=MGYuAAAAYAAJ&dq=editions:v9KYeuq_PbgC&pg=PR3&redir_esc=y&hl=en#v=o nepage&q&f=true]. Bernal, J.D. (1939) The Social Function of Science, Routledge Blanqué, P (2014): Essays in Positive Fund Management, Economica

FUND MANAGEMENT STRATEGY

BERNSTEIN

42

23 August 2016

Bolla, Kohler and Wittig (2016): Index-Linked Investing—A Curse for the Stability of Financial Markets around the Globe? Journal of Portfolio Management Spring 2016, Vol. 42, No. 3: pp. 26–43 Brogaard, Ringgenberg and Sovich (2016): The Real Impact of Passive Investing in Financial Markets. Washington University in St Louis Working Paper No 15/6 Clark (2007): A Farewell to Alms, Princeton University Press DALBAR. (2015): Quantitative Analysis of Investor Behavior available at http://www.dalbar.com/ProductsampServices/AdvisorsSolutions/QAIB/tabid/214/Default.aspx Department for Communities and Local Government (2013): LGPS Structure Analysis available at https://www.gov.uk/government/uploads/system/.../Hymans_Robertson_report.pdf EU commission staff working document (2015): Action Plan on Building a Capital Market Union available at ec.europa.eu/finance/capital-markets-union/docs/building-cmu-action-plan_en.pdf Goyal A and Wahal S (2008): The Selection and Termination of Investment Management Firms by Plan Sponsors, The Journal of Finance, Volume 63, Issue 4 Grinold, R. (1989) The Fundamental Law of Active Management The Journal of Portfolio Management, vol. 15, no. 3 (Spring): 30–38. Haldane (2010): Patience and Finance. Speech at the Oxford China Business Forum, Beijing available at www.bankofengland.co.uk/ Kay (2012): The Kay review of UK equity Markets and Long-Term Decisions Making, Final Report available at https://www.gov.uk/.../bis-12-917-kay-review-of-equity-markets-final-report.pdf Kinnel (2016): Mind the Gap 2016, Morningstar Manager Research available at https://corporate1.morningstar.com/ResearchArticle.aspx?documentId=756760 Landes (1998): The Wealth and Poverty of Nations Levine and Zervos (1998): Stock Markets, Banks, and Economic Growth: Do well-functioning stock markets and banks promote long-run economic growth? worldbank.org Lieppold, Su and Ziegler (2015) Do Index Futures and ETFs affect Stock Return Correlations? Available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2620955 Marx, K (1867): Capital: Critique of Political Economy (Das Kapital, Kritik der politischen Ökonomie) Mikkelsen (2016): European Fund Expenses Are Decreasing in Percentage, But investors pay more in nominal values. MorningStar 2016 available at http://media.morningstar.com/uk%5CMEDIA%5CResearch_Paper%5C2016_Morningstar_European_Cost_Study_1708 2016.pdf Philosophical Economics (2016a): Index Investing Makes Markets and Economies More Efficient, http://www.philosophicaleconomics.com/ Philosophical Economics (2016b): The value of active management: a journey into indexville, http://www.philosophicaleconomics.com/ Piketty (2014): Capital in the Twenty First Century, Harvard University Press Rajan and Zingales (1996): Financial Dependence and Growth, NBER Working Paper 5758 Solow (2007): The Survival of the Richest? New York Review of Books, November 22, 2007 Skousen (1990) The Structure of Production

FUND MANAGEMENT STRATEGY

BERNSTEIN

43

23 August 2016

Sullivan and Xiong (2012): How Index Trading Increases Market Vulnerability, Financial Analysts Journal, March/April 2012 | Vol. 68 | No. 2 US Department of Labor (2016): Fact Sheet: Department of Labor Proposes Rule to Address Conflicts of Interest in Retirement Advice available at https://www.dol.gov/ebsa/newsroom/fsconflictsofinterest.html Vayanos and Woolley (2016): The Curse of Benchmarks, London School of Economics Financial Markets Group Discussion Paper No 747 Wermers and Yao (2010): Active vs passive Investing and the Efficiency of Individual Stock Prices, SSRN Wurgler (1999): Financial Markets and the Allocation of Capital, Journal of Financial Economics 58,187-214 Wurgler (2010): On the Economic Consequences of Index-Linked Investing in Challenges to Business in the Twenty-First Century

FUND MANAGEMENT STRATEGY

BERNSTEIN

44

Disclosure Appendix SRO REQUIRED DISCLOSURES •

References to "Bernstein" relate to Sanford C. Bernstein & Co., LLC, Sanford C. Bernstein Limited, Sanford C. Bernstein (Hong Kong) Limited 盛博香 港有限公司, and Sanford C. Bernstein (business registration number 53193989L), a unit of AllianceBernstein (Singapore) Ltd. which is a licensed entity under the Securities and Futures Act and registered with Company Registration No. 199703364C, collectively.

•

Bernstein analysts are compensated based on aggregate contributions to the research franchise as measured by account penetration, productivity and proactivity of investment ideas. No analysts are compensated based on performance in, or contributions to, generating investment banking revenues.

OTHER DISCLOSURES A price movement of a security which may be temporary will not necessarily trigger a recommendation change. Bernstein will advise as and when coverage of securities commences and ceases. Bernstein has no policy or standard as to the frequency of any updates or changes to its coverage policies. Although the definition and application of these methods are based on generally accepted industry practices and models, please note that there is a range of reasonable variations within these models. The application of models typically depends on forecasts of a range of economic variables, which may include, but not limited to, interest rates, exchange rates, earnings, cash flows and risk factors that are subject to uncertainty and also may change over time. Any valuation is dependent upon the subjective opinion of the analysts carrying out this valuation. Bernstein produces a number of different types of research product including, among others, fundamental analysis and quantitative analysis. Recommendations contained within one type of research product may differ from recommendations contained within other types of research product, whether as a result of differing time horizons, methodologies or otherwise. This document may not be passed on to any person in the United Kingdom (i) who is a retail client (ii) unless that person or entity qualifies as an authorised person or exempt person within the meaning of section 19 of the UK Financial Services and Markets Act 2000 (the "Act"), or qualifies as a person to whom the financial promotion restriction imposed by the Act does not apply by virtue of the Financial Services and Markets Act 2000 (Financial Promotion) Order 2005, or is a person classified as an "professional client" for the purposes of the Conduct of Business Rules of the Financial Conduct Authority.

To our readers in the United States: Sanford C. Bernstein & Co., LLC is distributing this publication in the United States and accepts responsibility for its contents. Any U.S. person receiving this publication and wishing to effect securities transactions in any security discussed herein should do so only through Sanford C. Bernstein & Co., LLC. To our readers in the United Kingdom: This publication has been issued or approved for issue in the United Kingdom by Sanford C. Bernstein Limited, authorised and regulated by the Financial Conduct Authority and located at 50 Berkeley Street, London W1J 8SB, +44 (0)20-7170-5000. To our readers in member states of the EEA: This publication is being distributed in the EEA by Sanford C. Bernstein Limited, which is authorised and regulated in the United Kingdom by the Financial Conduct Authority and holds a passport under the Markets in Financial Instruments Directive. To our readers in Hong Kong: This publication is being distributed in Hong Kong by Sanford C. Bernstein (Hong Kong) Limited 盛博香港有限公司, which is licensed and regulated by the Hong Kong Securities and Futures Commission (Central Entity No. AXC846). This publication is solely for professional investors only, as defined in the Securities and Futures Ordinance (Cap. 571). To our readers in Singapore: This publication is being distributed in Singapore by Sanford C. Bernstein, a unit of AllianceBernstein (Singapore) Ltd., only to accredited investors or institutional investors, as defined in the Securities and Futures Act (Chapter 289). Recipients in Singapore should contact AllianceBernstein (Singapore) Ltd. in respect of matters arising from, or in connection with, this publication. AllianceBernstein (Singapore) Ltd. is a licensed entity under the Securities and Futures Act and registered with Company Registration No. 199703364C. It is regulated by the Monetary Authority of Singapore and located at One Raffles Quay, #27-11 South Tower, Singapore 048583, +65-62304600. The business name "Bernstein" is registered under business registration number 53193989L. To our readers in the People’s Republic of China: The securities referred to in this document are not being offered or sold and may not be offered or sold, directly or indirectly, in the People's Republic of China (for such purposes, not including the Hong Kong and Macau Special Administrative Regions or Taiwan), except as permitted by the securities laws of the People’s Republic of China. To our readers in Japan: This document is not delivered to you for marketing purposes, and any information provided herein should not be construed as a recommendation, solicitation or offer to buy or sell any securities or related financial products. For the institutional client readers in Japan who have been granted access to the Bernstein website by Daiwa Securities Group Inc. (“Daiwa”), your access to this document should not be construed as meaning that Sanford C Bernstein is providing you with investment advice for any purposes. Whilst Sanford C Bernstein has prepared this document, your relationship is, and will remain with, Daiwa, and Sanford C Bernstein has neither any contractual relationship with you nor any obligations towards you.

To our readers in Australia: Sanford C. Bernstein & Co., LLC, Sanford C. Bernstein Limited and Sanford C. Bernstein (Hong Kong) Limited 盛博香港有限公 司 are exempt from the requirement to hold an Australian financial services licence under the Corporations Act 2001 in respect of the provision of the following financial services to wholesale clients: •

providing financial product advice;

•

dealing in a financial product;

•

making a market for a financial product; and

•

providing a custodial or depository service.

Sanford C. Bernstein & Co., LLC., Sanford C. Bernstein Limited, Sanford C. Bernstein (Hong Kong) Limited 盛博香港有限公司 and AllianceBernstein (Singapore) Ltd. are regulated by, respectively, the Securities and Exchange Commission under U.S. laws, by the Financial Conduct Authority under U.K. laws, by the Hong Kong Securities and Futures Commission under Hong Kong laws, and by the Monetary Authority of Singapore under Singapore laws, all of which differ from Australian laws. One or more of the officers, directors, or employees of Sanford C. Bernstein & Co., LLC, Sanford C. Bernstein Limited, Sanford C. Bernstein (Hong Kong) Limited 盛博香港有限公司, Sanford C. Bernstein (business registration number 53193989L) , a unit of AllianceBernstein (Singapore) Ltd. which is a licensed entity under the Securities and Futures Act and registered with Company Registration No. 199703364C, and/or their affiliates may at any time hold, increase or decrease positions in securities of any company mentioned herein. Bernstein or its affiliates may provide investment management or other services to the pension or profit sharing plans, or employees of any company mentioned herein, and may give advice to others as to investments in such companies. These entities may effect transactions that are similar to or different from those recommended herein. Bernstein Research Publications are disseminated to our customers through posting on the firm's password protected website, www.bernsteinresearch.com. Additionally, Bernstein Research Publications are available through email, postal mail and commercial research portals. If you wish to alter your current distribution method, please contact your salesperson for details. Bernstein and/or its affiliates do and seek to do business with companies covered in its research publications. As a result, investors should be aware that Bernstein and/or its affiliates may have a conflict of interest that could affect the objectivity of this publication. Investors should consider this publication as only a single factor in making their investment decisions. This publication has been published and distributed in accordance with Bernstein's policy for management of conflicts of interest in investment research, a copy of which is available from Sanford C. Bernstein & Co., LLC, Director of Compliance, 1345 Avenue of the Americas, New York, N.Y. 10105, Sanford C. Bernstein Limited, Director of Compliance, 50 Berkeley Street, London W1J 8SB, United Kingdom, or Sanford C. Bernstein (Hong Kong) Limited 盛博香 港有限公司, Director of Compliance, Suites 3206-11, 32/F, One International Finance Centre, 1 Harbour View Street, Central, Hong Kong, or Sanford C. Bernstein (business registration number 53193989L) , a unit of AllianceBernstein (Singapore) Ltd. which is a licensed entity under the Securities and Futures Act and registered with Company Registration No. 199703364C, Director of Compliance, 30 Cecil Street, #28-08 Prudential Tower, Singapore 049712. Additional disclosures and information regarding Bernstein's business are available on our website www.bernsteinresearch.com.

CERTIFICATIONS •

I/(we), Mark Diver, Inigo Fraser-Jenkins, Paul Gait, Alla Harmsworth, Sarah McCarthy, CFA, Senior Analyst(s)/Analyst(s), certify that all of the views expressed in this publication accurately reflect my/(our) personal views about any and all of the subject securities or issuers and that no part of my/(our) compensation was, is, or will be, directly or indirectly, related to the specific recommendations or views in this publication.

Approved By: RP Copyright 2016, Sanford C. Bernstein & Co., LLC, Sanford C. Bernstein Limited, Sanford C. Bernstein (Hong Kong) Limited 盛博香港有限公司, and AllianceBernstein (Singapore) Ltd., subsidiaries of AllianceBernstein L.P. ~1345 Avenue of the Americas ~ NY, NY 10105 ~212/756-4400. All rights reserved. This publication is not directed to, or intended for distribution to or use by, any person or entity who is a citizen or resident of, or located in any locality, state, country or other jurisdiction where such distribution, publication, availability or use would be contrary to law or regulation or which would subject Bernstein or any of their subsidiaries or affiliates to any registration or licensing requirement within such jurisdiction. This publication is based upon public sources we believe to be reliable, but no representation is made by us that the publication is accurate or complete. We do not undertake to advise you of any change in the reported information or in the opinions herein. This publication was prepared and issued by Bernstein for distribution to eligible counterparties or professional clients. This publication is not an offer to buy or sell any security, and it does not constitute investment, legal or tax advice. The investments referred to herein may not be suitable for you. Investors must make their own investment decisions in consultation with their professional advisors in light of their specific circumstances. The value of investments may fluctuate, and investments that are denominated in foreign currencies may fluctuate in value as a result of exposure to exchange rate movements. Information about past performance of an investment is not necessarily a guide to, indicator of, or assurance of, future performance.