Bond Vigilantes: time to rustle QE out of town

• Quantitative easing (QE) has pushed asset prices up by relentlessly pushing returns for investors down. If QE is reversed, returns should go back up – meaning asset prices should fall.
• That prospect seems to be worrying many media commentators and fund managers. Yet share prices have shrugged off their concerns – until now. This post tries to understand why this is the case – and why it may change.
• Risk free rates of return (i.e. the yields on government bonds) in an economy tend to approximate to the GDP growth of that economy. QE was introduced in developed markets because of low economic growth. The low returns and resulting high asset prices we are experiencing may therefore be fully justified by the structurally low GDP growth of developed economies – even after QE is withdrawn.
• However, returns on risky assets like equities have another important component, which sits on top of the risk free rate: the risk premium. QE has done the same thing to risk premiums as it has to risk free rates, suppressing them to ultra-low levels. Investors could therefore get a higher return on their investments in risky assets if risk premiums rise from their QE induced lows – despite risk free rates being capped by low economic growth.
• Worryingly, there are many reasons for risk premiums to rise from their current lows. Global cross-border bank debt has increased more than fourfold since 2001 and money supply to GDP has reached unprecedented levels. This created a buyer’s market – where too much money was chasing assets – and drove risk premiums down. Those supportive factors will reverse as QE unwinds. Demographics, which previously benefited asset owners, will soon begin to favor asset buyers. The risk premium is there to reflect real risks (the clue is in the name) and the environment for big global corporations’ profits and big governments’ debts has never been better. It can only get worse from here – which is a big risk. Finally, asset prices are so expensive that you need to be phenomenally rich to retire with dignity – that is not sustainable.
• Rising rates of return for investors are likely to result in a comeback for a long forgotten investment legend: the bond vigilante who, once upon a time, drove down the prices of bonds to enforce discipline on the governments issuing them. Far from causing any economic crisis, this will allow rational capital allocation to resume and economic growth to be driven by innovation, creativity and other sustainable factors – rather than by a constant increase in debt.

 

Expensive markets

Asset markets have been strong since 2009. Despite the recent sell off, the S&P500 index is trading close to all time highs:

That’s largely because quantitative easing (QE), in other words the expansion of central bank balance sheets to purchase government bonds and assorted securities held by commercial banks, has pushed the price of those government bonds and other securities up, with a knock-on effect on prices of all other asset classes (mainly equity, infrastructure and real estate). The following chart, published in zerohedge, shows the unprecedented extent of this central bank balance sheet expansion:

With this amount of central bank money having been pumped into asset markets, and with those markets now so expensive (as we shall see below), I think anyone buying the market at these elevated levels is at risk if QE is abandoned (assuming it ever is, of course).

That risk can be broken down into two separate components, one minor, one major.

 

QE winners and losers

The lesser risk is the economic impact of the end of QE, on corporate cash flows and government tax income. The risk is minor because, I would argue, QE had a muted impact on the broad economy. That’s because it was transmitted to the real economy through the banking sector, which was (and still is) rebuilding its equity to meet new, stricter capital requirements introduced in the wake of the global financial crisis (GFC), this at a time when low interest rates (brought about by QE) were reducing the sector’s profitability (mainly by depressing the return on banks’ deposits and float). The extent to which the capital the banks received due to QE was used to to lend to individuals or businesses was limited by the banks’ attempts to reduce the size of their balance sheets, including loans, relative to their equity and by the slow growth of that equity due to the impact of QE on their profitability. The new capital requirements to which the banks were reacting also attached a higher capital charge to corporate, particularly SME lending. This encouraged banks to use most of the QE capital they injected into the economy (rather than retaining it to bolster equity) to make safe, mainly property-related loans, whose main impact was to boost the price of real estate and other assets, not economic activity.

During this period, capital allocation to corporates in general, whether of debt or equity, has, understandably, been concentrated in a narrow sector of economic winners, from digital powerhouses such as Amazon, Facebook and Google to low cost category killers like Lidl, TK Maxx or Cote Restaurants (referred to hereafter as the “QE winners”), which have used it to expand capacity and eat into the returns of weaker entities (the “QE losers,” e.g. advertising agencies, newspapers, high street retail, conventional supermarkets or independent restaurants) who – their weakness notwithstanding – still constitute the majority of participants and employ the majority of people in their respective sectors. This, in turn, encouraged banks and other investors (such as venture capitalists) to shun those weaker players, creating a vicious (or virtuous, depending on your perspective of course) circle in which the allocation of capital to a narrow group of economic winners enabled them to outspend and thereby further consolidate their position vis à vis their weaker rivals and in turn encouraged investors to give them more capital …

At the same time, cheap borrowing made available by QE is probably allowing some of the weak “QE loser” businesses to limp along and these may fail if QE is withdrawn. But that, as elaborated further below, is not necessarily bad for the economy – since better businesses will tend to take their place.

Of course, QE also allowed governments to finance their borrowing cheaply and keep spending, which, in normal circumstances, should support the economy with conventional Keynesian stimulus. However, QE has in the main enabled government spending to remain stable or to be cut less in key areas, despite a reduced or slowly growing tax take, while certain elements of spending related to an ageing population continued their relentless rise. Therefore, although talk of austerity is odd in a context of record government debt and spending, QE did not provide much in the way of a Keynesian demand support to the economy, as has indeed been noted by many advocates of such Keynesian measures (such as Paul Krugman or Ann Pettifor, cited below). In any case, a significant proportion of the government spending enabled by QE has been allocated to highly paid civil servants (e.g. in the US, UK or EU) or virtue signalling and highly paid diversity officers. Opinions may vary on the merits of these investments, but none can be claimed to have much of a positive impact on economic growth per se. Meanwhile, genuine infrastructure spending has been concentrated on expensive vanity projects such as the UK’s HS2 rail project.

This, paradoxically, means that the economy does not have much to lose from a withdrawal of QE – because the broader economy never really gained that much from it. This point will be developed further below when we examine why a fall in asset markets does not necessarily entail the economic collapse which currently seems to be predicted by many.

 

What is a discount rate?

The greater risk is not the economic risk of a fall in corporate cash flows or tax income, but the valuation risk of an increase in the discount rate applied to those corporate cash flows, or to the government bond coupon payments funded by that tax income. “Discount rate” is another way of describing the rate of return. Since most people intuitively understand what rate of return means, it provides a useful starting point for explaining the crucial concept of the discount rate (if you’re already familiar with discount rates, as most financial analysts are, then you can skip this section).

If an asset pays you $3m a year and is priced at $100m, what is its rate of return? The answer is $3m return divided by $100m cost of investment = a 3% return on that investment. So far, so obvious. We refer to the discount rate when we do the same calculation in reverse and attempt to put a price on a future annual cash flow stream of $3m. To break that down in simple terms, for anyone who isn’t used to this sort of calculation, we can start by looking at the value of the first of those annual cash flows, i.e. the first $3m we receive in a year’s time. $3m today is worth … $3m. That’s obvious. To calculate the value of $3m in a year’s time, you simply discount it back at your discount rate of 3%, so it is worth $3m/[1 + 3%] = $2.91m. To some this may seem a little abstract. A common sense way of explaining it is that if you start with $2.91m today and earn 3% on it, it will be worth $3m in one year’s time: $2.91m x 3% = $0.09m cash flow from your investment, add that to your initial $2.91m and you end up with … $3m. Accordingly, if your required rate of return is 3%, $2.91m today will be worth exactly as much to you as $3m in a year’s time. Therefore, to do the same in reverse and calculate what $3m in a year’s time is worth today, you just discount it back at your required rate of return. If that doesn’t make sense don’t be embarassed to tweet me @semperfidem2004 (or @kafkascastle1 while @semper is banned).

Now, to calculate the value of an investment which pays you $3m every year in perpetuity, you have to discount next year’s $3m at 3%, the $3m you get in two years at 3% compounded over two years (6.1%), the $3m you get in three years at 3% compounded over three years and so on all the way out to perpetuity … and then add all of those discounted values up. That is an onerous calculation, but, luckily, the financial algebra gurus worked out that you can simplify it to asset value = perpetual income stream ($3m) divided by the required rate of return or discount rate (3%). In this example, a $3m perpetual income stream discounted at 3% = $3m/3% = $100m. And that makes intuitive sense: if $3m return on $100m investment gives you a 3% rate of return, then, by definition if you expect a 3% rate of return, an investment that pays $3m per year will be worth $100m to you.

 

What happens when discount rates go up?

What does this mean for QE? It’s well known that QE means central banks printing money to buy bonds. What’s less well understood is that, in so doing, those central banks dramatically reduced the discount rate we just defined. That is because they weren’t (and still aren’t) buying the bonds to get a good return. They were buying them to finance government deficits at an affordable rate for the government. They are driven by political, not investment considerations. What is unusual – even bizarre – here is that, in these transactions, government bond yields are not determined by the appetite for yield expressed by different investors buying and selling in the market, but by a single, politically motivated buyer who unilaterally sets the price. It is institutional price fixing of government debt. Crucially, the central banks used their price fixing discretion to dictate that interest rates – in other words the discount rate – on that debt should be artificially low. The justification given for QE’s institutional price fixing was that it has enabled governments to survive without making drastic cuts to expenditure and prevented the banking system from freezing up, thereby keeping economies from collapsing.

However, the impact of QE was not limited to its first order impact on government borrowing costs. Faced with low yields on government bonds and lower yields on the other assets purchased from commercial banks funded by QE, investors turned to the rest of the market to find better yields. That in turn pushed down the yield on other assets like equities or real estate. By helping governments to keep on spending, QE ended up dramatically reducing the discount rate across all asset classes. Now, the mathematical illustration we saw above was based on a common sense principle: the more expensive an investment is, the lower its return. Therefore the inevitable, algebraic result of reducing the discount rate is to make assets more expensive. In this particular case, the reduction of borrowing rates to help governments keep spending and banks keep lending ended up making a handful of big asset owners very, very rich. If you want to understand why the 1% are so rich today, you should start by examining the impact of QE on asset valuations.

The flip side of this principle is that any rise in discount rates from here, due to an abandonment or reversal of QE, will automatically lead to corporate assets and government bonds getting cheaper. In other words it will lead to a fall in the price of those assets, all else being equal.

To return to our example above, we saw that the value of a perpetual income stream of $3m discounted at 3% (which, until recently, was the yield to maturity on 10 year US treasuries; the current yield, after the increase which led to the recent stock market sell-off, is 3.2%) is $100m ($3m/3%). If you increase that discount rate by a mere 0.5% from those low levels, the value of the perpetual income stream – all else being equal – falls by almost 15% to $85.7m ($3m/3.5%). If we approach what were once normal discount rates of 5% (before QE pushed them down), then the value of that income stream goes down by 40% to $60m ($3m/5%). In short, we only need interest rates to go up by a few percentage points for asset values to halve – just because of the discount rate. Of course, the stated aim of central banks is to only raise rates in line with any resumption in growth, which would, they hope, compensate for the rise in discount rates. Whether this can offer any comfort to investors will be examined below.

 

The market seems to be relaxed about that risk

Yet markets, be they equity or bond or other, seem oddly unperturbed. The sell-off in US equities earlier in the year left the S&P500 index only 7.5% off what was then an all-time peak reached in January, and the market subsequently rebounded to all-time highs which are a full 85% above the peak reached in 2007 before the GFC. This is without counting dividends received by investors in that period, which, compounded, are worth around another 24%, taking total shareholder return over 100% in that period. Markets are not infallible, but they are powerful mechanisms for anticipating the future and discovering value; they are the best representation we have of the wisdom of crowds. The outcome of thousands of traders, investors and algorithms buying and selling usually reflects the reality of a situation long before any one of those individual traders or investors (or maybe today even AI algorithms) has been able to rationalise it; the rationalisation usually comes after the market move – particularly at inflection points. Therefore, although the market may sometimes be mad, as it was before the GFC or during the dotcom bubble, it is always worth trying to understand what its collective wisdom may be discerning or anticipating at any given time.

In any case, even when the market is mad, the rationalisations of most individual market participants at the time are far crazier than the market’s and indeed often verge on gibbering insanity – to which the market’s more moderate, aggregate insanity is infinitely preferable (this is of course entirely compatible with there always being a few lone voices of sanity, such as Robert Schiller in the run-up to the dotcom bubble, which correctly see through the market’s madness). Many press commentators are currently bearish, as are many active asset managers, who, according to some surveys, are holding high levels of cash. This of itself is reassuring: such commentators are typically contrarian indicators, so, if they’re pessimistic, it’s a good rule of thumb to expect markets to go up. Indeed, markets are less likely to fall in such a scenario, where there is money waiting to buy the dip, than one in which everyone is bullish and fully invested. It is therefore worthwhile exploring what the market’s strength may be reflecting and which nervous active managers or journalists haven’t (yet) discerned.

 

Low growth = low yields?

I think that what the market’s current strength is reflecting is the impact of the new post-GFC low growth environment on asset valuations. QE was introduced, in response to the GFC, as an attempt to prevent growth from tipping below a critical level; it is a policy response to the threat of low – or even negative – growth. Therefore, if there really is a QE asset bubble, then the prices of assets in that bubble may somehow be connected to the low growth environment which is offered as justification for QE. In other words, it is possible that it is this low growth environment which the market is reflecting in current asset price valuations. That, at least, is the hypothesis which will be explored in this section.

Our exploration of that hypothesis starts with the acknowledgement that there is a commonly acknowledged theoretical relationship between a country’s GDP growth and its risk-free interest rates and that this theoretical relationship can actually be observed in the data, as in this chart from the St Louis Fed:

There are good reasons for this relationship to hold. One reason is that if it didn’t an arbitrage would arise. This is best explained by drawing an analogy with the relationship between corporate cost of capital, of which risk-free interest rates are a component, and growth at the individual company level. For a company, potential growth in invested capital, absent capital injections or leverage – in other words organic growth in invested capital – equals return on capital, compounded. If return on capital is stable, then organic growth in corporate cash flow also equals return on capital (assuming all capital is reinvested, which is not always the case as we will discuss below). If, in contradiction to this rule, cost of capital in an industry were ever to go below return on capital, then investors could earn an excess return (equal to the spread between return on capital and its cost) from injecting capital into that industry. That spread would encourage investors to inject capital into the industry. This injection should, eventually, drive down returns (both by adding to the capital base in the denominator and reducing prices through increased supply) until return on capital equalled cost of capital. Vice versa, if returns on capital were ever to go below cost of capital, no capital would be injected into the industry and, over time, capital would be allowed to deplete or taken out of service, reducing capital invested until it eventually drove up returns to the cost of capital [1].

In short, for companies, potential organic growth in corporate cash flow should roughly equal return on capital. Therefore, over a certain period of time and absent distorting factors, one would expect companies’ real cost of capital to equal potential organic growth in their economic output (for which cash flow is a proxy), just as GDP growth gravitates around 10 year bond yields in the chart above. However, before looking at how this relationship between companies’ growth rates and cost of capital can be extrapolated to whole economies, we need to look at the frequent and persistent divergence that can be observed between those two metrics.

First, there are obviously leads and lags between the two, as can easily be seen in the chart of US 10 year yields and GDP growth above. This is partly due to the fact that GDP growth is released quarterly and can swing up or down due to temporary factors. Long term bond yields tend to be more stable because they “look through” the statistical “noise” of these short term fluctuations in GDP. At the company level, divergences between cost of capital and return on capital are also due to the fact that cost of capital only tends to equate to prospective returns on capital and realised returns may not correspond to those expected when an investment is made. In particular, we see industries with many years of returns above (below) cost of capital as injections (reductions) in invested capital are not rapid enough to keep pace with secular growth (obsolescence) in the industry. Despite these leads and lags, over time, we could expect cost of capital to gravitate around return on capital and therefore around sustainable growth in economic activity for individual companies, for the reasons given above. This can also be seen in the chart above, in which, for example, yields take a while to catch up with the inflation driven growth in GDP at the end of the seventies, with both series finally synchronising in 1983, after the recession induced in part by the taming of inflation.

Second, beyond those timing issues, we can observe that in the last thirty years, at least, developed market quoted companies’ cost of capital has been less than their return on capital. As we will see in more detail below, this means, by definition, that the developed market corporate sector has typically traded above, and for the most part significantly above, invested capital or book value. The accounting term for the premium of a company’s market value over its asset value is “goodwill.” This means that the axiom that competitor capital is inevitably attracted into an industry due to this excess return (i.e. return on capital being greater than cost of capital) has not held good in all cases or indeed most cases in recent history. The reason is that, in many industries, dominant market share, brand value, know-how and other advantages prevent new entrants from gaining share in the industry and thereby driving down the returns of the incumbents. Such industries are therefore dominated by a small group of players earning a higher return on capital than their cost of capital and therefore making an “excess return.” These barriers to entry and “competitive moats” (to use a term frequently employed by Warren Buffet in his annual letters to Berkshire Hathaway shareholders) enable some, indeed many companies’ return on capital to resist the gravitational pull toward their cost of capital which was described above [2].

 

Re-investment rates: from individual companies to whole economies

We noted that not all companies reinvest all their cash flow and that only a company which reinvests all of its capital will grow at its return on capital. A company which reinvests none, and effectively pays all its cash flow out and operates as a cash cow, cannot be expected to grow in real terms over the long term. It is likely that the higher the price to book of a sector, in aggregate, the more dominated it will be by a few leading companies – otherwise you would find other companies investing in that sector to exploit the excess returns available there. This means there will tend to be fewer re-investment opportunities in that sector and therefore, in general, companies in that sector won’t be able or find it attractive to reinvest their cash in it. This is not always the case, however: one can also frequently observe companies and industries with high barriers to entry and trading significantly above book value which re-invest all (or more than all when they borrow) of their cash flow. A good example of this is consumer goods companies which use their dominant position in one part of the market to expand into another, adjacent one. The relationship between reinvestment rate and goodwill premium is not linear.

At the macro level, however, the capital from some of the dominant companies trading above book and lacking re-investment opportunities, or mature companies which do not grow and therefore don’t reinvest, is typically recycled into the faster growing, capital hungry companies. Therefore, the organic growth in the economic activity of the corporate sector in aggregate will tend to approximate more closely to its return on capital than the growth of any individual company does to its return on capital.

In short, the corporate sector can trade above book value to a certain, limited extent, due to barriers to entry, and it reinvests its cash flow to different extents. The principles analysed above can, with these caveats, be scaled up beyond the corporate sector to the whole economy and its treasury. In any economy, a certain portion of output is reinvested and the rest is consumed. The real return on the whole economy’s reinvested capital determines that economy’s real organic growth rate in output (i.e. absent increased financial leverage, whether domestic or external).

Therefore, an analogous arbitrage argument to the one we discussed in connection with individual companies and the corporate sector applies to the economy as a whole and to government debt in particular. A government that could borrow at interest rates below its growth rate could use this to issue debt, invest the proceeds – or better still allow a portion of the proceeds to be invested by the private sector (via tax breaks for example) – at a higher rate than the cost of that debt, and thereby generate GDP growth above the cost of that debt. In such a scenario, it would – in theory – make sense for the government to take on more debt in order to benefit from this spread. The problems with this argument, which is currently popular in Left leaning circles, are that government investments, e.g. in infrastructure, carry a much higher risk than the government debt itself, that the positive impact of government investment on growth is highly uncertain, particularly in mature economies, and that increasing debt causes an artificial but unsustainable increase in demand which is usually followed by a recession (I analyse that problem in depth here). Even absent those problems, the more the government takes advantage of this arbitrage by issuing debt, the more one could expect the increased issuance (supply) of debt to force down (up) the price of (yield on) that debt, until it eventually reaches the country’s growth rate. This is part of what is driving the relationship between GDP growth and cost of debt in the chart above. Of course, the whole point of QE is to make the cost of government debt independent of supply and therefore break this link between GDP growth and cost of debt.

GDP growth is not only a theoretical floor for real yields but also a theoretical cap. In simple terms, if you add $100bn to a country’s debt and the interest rate on that debt is 5%, you add $5bn to its annual interest payments. That interest is paid by tax income, which is raised from the economy, whether individuals or businesses. Unless economic output increases by 5%, then you have a worsening of the ratio between economic output, which is the total amount available to service debt, and interest payments, in other words you have a decline in the country’s solvency. If, eventually, debt reaches 100% of GDP, then any increase in debt, if such parameters obtain, will result in a decline in GDP after interest payments. In other words, if a country is growing at a lower rate than its cost of debt, then, like Greece, it will eventually be unable to service any increase in that debt. Of course, QE by definition meant that the recent massively increased supply of government debt had no impact on its price, given that demand for that debt could be increased simply by printing money. The axiom we are discussing only holds in a world without QE.

In conclusion, the faster (slower) an economy grows the more (less) demand it has for capital and therefore the higher (lower) its cost of borrowing. Therefore, for government bonds, which reflect the aggregate economics of a country and define its risk-free return, it is rational to expect sovereign yields to approximate GDP growth – unless you have a distorting factor like QE. QED.

 

Low growth = high valuations

The paradoxical implication of this axiom is that the lower the sustainable growth in any market, the higher its valuation should be. This may seem to run counter to the common intuition that investors are inclined to “pay a premium for growth.” But that rule only applies to investments from which investors derive a benefit from growth in the cash flows generated by the asset in which they invest, for example equities or real estate. Even for such investments, the growth rate is deducted from the cost of capital when setting a theoretical value for the growing asset; it doesn’t influence the cost of capital per se. Growth therefore does not make you pay more for bonds, either corporate or government. Of course, anticipated future growth in cash flows should be reflected in the creditworthiness of the entity issuing the debt. But with governments we are dealing with the risk-free rate and therefore that consideration can be left to one side, for now.

According to this axiomatic reasoning, if future growth rates remain structurally low after QE is abandoned, as implied by the purported necessity for QE and indeed emphasised by its (perverse) effects on capital allocation, then, paradoxically, cost of capital should also remain low and asset valuations should remain expensive.

Indeed Japan, one of the lowest growth economies in the developed world, also has a structurally high PE stock market and ultra-low government bond yields. The market at the moment may therefore be correctly anticipating structurally lower growth and lower returns and, therefore, what currently seem like bubble valuations may be no more than a logical, even mechanical reflection of that low growth reality in a lower cost of capital. Equity fund managers (of all nationalities) who make a living from investing in Japanese equities are typically, in my experience, enthusiastic in expounding the gloomy notion that Japan, far from being an economic anomaly, represents the future for all economies. This of course makes them the life and soul of any party.

 

Risk premium

According to the axioms above, the real yield on government bonds can’t rise above the real GDP growth rate. Does that mean investors are therefore stuck with bad returns in a low growth environment? Not necessarily. Unlike government bonds, the returns on risky assets, such as equities or property, can rise, entirely consistently with the axioms above, by boosting the “risk premium.” The risk premium is simply the yield on a risky asset minus the risk-free rate and minus the expected growth in that asset’s cash flow; it is the yield premium you get for the risk associated with the cash flow of the company you invest in versus the cash flow of a risk-free government bond coupon.

In other words, the yield on equities could go up from, say, 3.9% to 5% while the risk-free rate stays at 3%, capped by the nominal growth rate of the economy. In such a case, the risk premium would have increased by 1.1% ([5% – 3%] – [3.9%-3%] = 2% – 0.9% = 1.1%). The risk premium implied by current market valuations will be analysed in detail below. The investor in this example would get a better return, but without violating the law that risk free interest rates can’t go above GDP growth. In short, investors in risky assets could earn a higher yield simply by raising the risk premium they demand from the assets they invest in.

 

Rising returns need not be fatal

Returns can, in the current investment environment, rise on risky assets, increasing remuneration for investors, without rising above the critical level of corporate return on capital. Why does this matter? Because if cost of capital goes above return on capital it can cause serious problems, resulting in capital reduction (as described above) and even bankruptcy. Just as it does for governments when risk free rates go above a country’s GDP growth. In the current environment, however, returns on risky assets can comfortably go up for the investor without hitting the critical levels above return on capital.

The reason we can raise the risk premium, and thereby returns for investors, without encountering the critical level where cost of capital goes above returns on capital, is that returns for investors are not low because corporate returns on invested capital are low – indeed, returns on capital are currently very high. Rather, investor returns are low because the equities through which investors gain access to companies’ underlying return on capital are trading at a high price to  that invested capital. The higher the price of an asset, the lower its returns. In other words, returns on equities are very low because of their price, not because of the returns on their underlying capital (the high price of equities to invested capital is due in large part to the spillover effect of QE from government bonds to other asset classes, as described above). This means that the cost of capital is so much lower than return on capital, at a safe distance so to speak, that risk premiums could rise a long way before they take cost of capital above the critical level where it overtakes return on capital. In other words, investors can get significantly better returns on their investments without taking cost of capital anywhere near that risky level.

In practice, this means that equity risk premiums can rise without equities trading below book value. Property yields, too, could rise, taking the price of property to build cost ratio closer to unity, but without making houses trade below their replacement cost – which would make housebuilding uneconomical. Similarly, corporate bond spreads could rise without taking them above a company’s return on capital, at which level the company would be unable to service its debts. That being said, current corporate spreads are so very low that any rise would clearly push a lot of highly leveraged companies over the edge, making corporate bonds particularly unattractive at present, in my view. High risk bonds, in which investors could lose 100% of their capital, currently offer a mere 3-4% yield premium to high quality bonds:

 

How low is the risk premium?

What are the actual returns on capital, costs of capital and prices to book encountered by investors in the current market? The first thing we need to know in order to calculate this is the cash flow yield on equities. We calculate this by dividing aggregate cash flow from, say, companies in the S&P500, by the market value of the S&P500.  According to the estimates of Terry Smith, manager of the successful Fundsmith global equity fund and author of a reference work on accounting scandals, Accounting for Growth, the cash flow yield on the companies in his portfolio, which have sustainable competitive advantages, high return on capital and strong balance sheets was 3.7% at the end of 2017, versus 3.9% in aggregate for stocks in the US S&P (whose chart is displayed above) and 5.6% in aggregate for the top 100 stocks in the UK (the Footsie). The higher yield on UK stocks is mostly a sector effect, with high price, low yielding global dominant companies like Facebook accounting for a big percentage of the S&P, and the Footsie being more dominated by oil companies whose profits may disappear because of electric cars, pharma companies facing patent cliff edges, banks whose earnings can be wiped out in the next financial crisis, and other troubled companies like Tesco and other retailers.

The relationship between these cash flow yields at current prices and the return on capital of the underlying business assets can be illustrated by imagining a company with the same metrics as those identified above for the market as a whole, i.e. a 15% return on invested capital (the aggregate return estimated by Smith for companies in the S&P, versus the 28% weighted return for the companies in his portfolio) and offering investors a 3.9% cash flow yield at current prices and so trading at a price to invested capital of 3.85x (15%/3.9%). If that company had invested capital of $1bn, an easy round number, its market value would equal $3.85bn (3.85 x $1bn) and it would generate $150m in annual cash flow (15% x $1bn). Accordingly, $150m cash flow divided by $3.85bn price gives the investor a 3.9% return. (QE)D. In other words, the higher the price/invested capital ratio, the lower the yield, by definition (yield = cash flow/price; return = cash flow/invested capital => yield = return ratio divided by the price/invested capital ratio).

Accordingly, equity investors could get a better return if they only, say, paid 3x price/invested capital rather than the 3.85x implied by year end 2017 valuations. That would result in a price of $3bn (3 x $1bn) and a yield of 5% ($150m/$3bn). This may not seem like a lot, but compounded over twenty years it results in an increase in the investor’s pot of assets equivalent to ~50% of their initial investment (and ~24% more than the value of their investments growing at the lower yield of 3.9%). Twenty years is not far from the average period during which a typical saver invests for retirement and few people would be indifferent to a 50% nominal increase in their pension pot. And, crucially, its price could still more than halve from there to $1bn before it started to trade at the critical level below book value. Therefore, looking at the current figures for the S&P, we can clearly see that there is plenty of scope for market returns to improve before that critical level.

We know risk free rates were 3% up until recently. Using Fundsmith’s figures we observed a cash flow yield on the S&P of 3.9% end 2017. But, as we explained above, that observed yield is composed of cost of capital and growth in cash flow (return on investment required by investors = cost of capital minus future cash flow growth); sustainable future cash flow growth is subtracted from cost of capital to arrive at the return on investment required by investors. Therefore, to calculate the cost of capital implied by the market yield, you need to estimate the future growth in cash flow. Here, as a very low-quality but high profile proxy for growth in cash flow is the inflation adjusted earnings per share of the S&P500, as calculated by S&P (the publishers of the index):

We can see here that EPS end 2017 was just over 10% above the previous peak of 2006, a compound growth of 0.9%. 1% real growth is not an overly pessimistic target for corporate cash flow in the future. Given the benefit to past growth from increased leverage it would be unwise to assume real growth above 1% if debt is stable or – heaven forfend – decreasing.

Whatever the case, nominal yields on 10 year treasuries are 3.2%. So nominal GDP growth implied by this bond yield is 3.2%. We can be generous to current valuations and take the 3.2% growth implied by the bond market as our reference point. This would assume either 2.2% long term inflation (3.2% bond yield minus the 1% real growth calculated above) or lower than 2.2% inflation combined with higher real growth than the 1% figure estimated in the previous paragraph. We can go on to compare current bond yields to current equity yields. US equities are up about 11% since 2017. If S&P cash flow is up 6% per share since (my estimate) then cash flow yield has gone from 3.9% at the end of 2017 to 3.7% now. So the S&P offers offers a nominal 3.7% cash flow return with an implied 3.2% growth, compared to treasury yields of 3.2%, implying a risk premium of 3.7% (3.7% + 3.2% growth – 3.2% risk free rate). 3.7% risk premium for the S&P relative to government bonds is one of the lowest post-GFC (the chart below doesn’t adjust for prospective earnings growth, which was much higher in the 87-07 period, forcing treasury above equity yields). Shares have, since around 2012, continued to get more expensive despite rising 10-year rates:

What does this mean for investors? It means the return the investor gets can be increased by increasing the risk premium applied to the company’s cash flows even if there is no growth at all in those cash flows. The hypothetical company analysed above may not be growing at all – despite having a return of 15% on invested capital – because, as we pointed out above, it is not necessarily reinvesting its cash flow to grow invested capital, but instead may be paying it all out to investors (in which case its capex would equal depreciation). Indeed, such “cash cow” behaviour should be more and more common if we really are in the low growth environment on which the post-QE valuations are predicated. The improved return through increased risk premium argument holds whether a company pays its cash flow out or reinvests it; if it reinvests its cash flow it should grow and that growth rate would, in theory (as described above), be deducted from its cost of capital when valuing the firm.

In summary, in a low growth world, investors in risky assets can get a higher return simply by demanding a higher risk premium – even if there is no growth in the cash flow of those assets. The current ultra-low levels of equity risk premium leave a lot of scope for investors to get an improved return via an increased risk premium without taking cost of capital above the critical level of return on capital. The ceiling for returns for investors in risky assets is, accordingly, not the economy’s growth rate but corporate return on capital. It is therefore possible for investors to get high returns in a low growth, Japan-like environment simply by paying less for assets, i.e. reducing the price to book they accept and increasing the risk premium they demand.

 

Buyer’s strike

We explained above how competitive advantage allowed companies to enjoy a cost of capital below their return on capital without seeing erosion of their returns by new entrants. What we did not discuss is what determines that cost of capital. The possibility we are envisaging in this post is a “buyer’s strike” which pushes the risk premium and therefore the cost of capital up, i.e. pushes returns up and valuations down. But why would investors suddenly demand a higher risk premium having accepted a lower one for all these years? The idea of investors having any say in the price of assets, to the point of actually forcing their prices down in a meaningful way, may seem quite outrageous, even insane, in the current environment of seemingly unlimited amounts of money chasing a finite number of investible assets.

The yield demanded from or offered on any asset is inherently subjective. You can never fully rationalise it or explain it. It is an example of what certain economists call “revealed preferences” (as articulated inter alia by Ken Binmore) and one of the metrics which markets are better at establishing, using the wisdom of crowds, than any economist is with their theories. We don’t know why the aggregate supply and demand for equities determines that they should offer a 3.9% prospective cash flow yield risk premium rather than 3.7% – or even 8.85% like in the good old days or 0.005% like in Japan. All we can see is the price actually paid for equities in the market at any given point, from which we can only infer the market’s required yield or – to adapt the economics term – “revealed yield preference.” We will now turn to the factors which might influence this metric.

The high price to invested capital of the current equity market reflects a high level of goodwill attributed to the market by investors, as described above. That is due to a number of reasons. First, a large part of the market in terms of market value consists of QE winners, who naturally attract a high valuation due to the strength of their business models. Also, as already mentioned, the low absolute level of returns on government bonds, often negative in real terms, has pushed investors to hunt for returns in riskier assets and therefore reduced (increased) the returns (prices) they are prepared to receive from (pay for) them. Clearly, both factors are intimately connected to QE.

That’s where we are now. In future, my hypothesis is that just as QE’s compression of risk free rates led to a hunt for yield that also compressed the risk premium and therefore the returns on risky stocks and pushed up the goodwill in the market, particularly in the dominant stocks which were beneficiaries of QE, so the end of QE should, according to the laws of gravity, result in an increase in the risk premium and a reduction in the goodwill attributed to the market, perhaps, particularly, to the small group of QE winners who previously led the market.

One could counterargue that the reason we have a low growth environment, despite the high returns on capital enjoyed by companies, is that very little of that return is reinvested. The fact that there is very little re-investment of capital implies that there is more available capital to invest, in the shape of the high cash returns generated by companies, than there are reinvestment opportunities. This is necessarily the case, so that counterargument runs, because if the capital were reinvested it would produce a return and therefore growth. If the capital is not being reinvested, according to this counterargument, it is being distributed to investors. Investors would therefore be sitting on a lot of cash. And if growth is low then one can also expect there to be few investment opportunities. If established corporates can’t find things to invest their cash in, then it’s unlikely start-ups or new entrants will. In such an environment, there would be too many investor dollars and not enough potential investments. Any attempt by one investor or group of investors to “play hardball” over prices would be defeated by other investors with cash “burning a hole in their pocket” who would be waiting for any fall in price to jump in. Since so many investors were hoping for such a fall in prices it would never happen.

This counterargument is flawed for a number of reasons. Firstly, because corporate cash flows are only a small contributor to the size of the monetary base and therefore their reinvested cash flow is only a small determinant of cash flow available for investment. A larger component, currently, is of course QE itself. As QE unwinds the amount of investment dollars waiting in the wings for valuations to fall will fall too. In addition to QE, the banking money multiplier expands the monetary base, according to classic economic theory and observation of narrow and broad money aggregates. Current banking debt levels, as measured by the BIS, has reached a vertigo inducing all-time high of $105 trillion (1.35x global GDP).

In the chart above we see bank debt almost quadrupled (as described in my blog on the GFC), from around $11.7trillion to $40.5 trillion from 2001, when China joined the WTO, to the pre-GFC peak of March 2008. Post-China’s WTO accession, it took six years for debt to grow by almost four times the amount it previously grew by in the 30 years of the post-Bretton Woods (1971) era recorded by the BIS.

Measuring what happened after that is complicated, as the gap in the chart above makes obvious. The growth in bank debt from 2001 to 2008, recorded by the BIS, happened at a time when the BIS only recorded cross-border debt and local debt in foreign currencies. Up to that point, local debt in local currencies didn’t change much and was not considered an important subject. It was therefore not recorded by the BIS. The BIS attributes its decision to begin collecting local debt in local currency data from the banks it monitored to the impact of the GFC on the financial system, reasoning that another financial crisis might arise in this area (the decision is explained here and here). The timing, however, coincides with significant expansion in central bank balance sheets, notably at that time by the Federal Reserve, which would have potentially been reflected in local debt in local currency; one could be forgiven for suspecting that the BIS might have been prompted to begin monitoring this data by the fact that QE caused a significant expansion in it. The maiden figure for local debt in local currency on March 2012, the first date on which it was recorded, was $42.8 trillion. It subsequently increased to $63.6 trillion by December 2013, when all the banks reporting to the BIS had filed their figures. If you compare the figure of $105 trillion in March 2018 to the previous peak in March 2008 plus the first time addition of the complete set of local debt in local currency of $63.6 trillion, summing to $104.5 trillion, you find that total bank debt is roughly back where it started post the previous March 2008 peak ($105 trillion March 2018 versus $104.5 trillion March 2008, adjusted for the first time addition of “local in local” debt from March 2012 to December 2013).

What we don’t know, of course, is how much local in local debt increased from 2008 to 2013. It probably did to some extent due to TARP and LTRO, but it looks as though QE has not done much more than take total bank debt back to just above its pre GFC peak. This is another reflection of the lack of impact of QE on the real economy, alluded to throughout this post. The fall in government bond rates and the expansion of the money supply due to QE did not result in an expansion in bank balance sheets to the benefit of the real economy. However, this is small comfort. The pre GFC peak was achieved by an incredibly rapid increase, representing 86.5% of 2001 world GDP.

QE was only able to maintain total bank debt at the nose bleed heights achieved pre GFC without really supporting the economy. However we can see an impact on the US M3 money supply (I use the monthly data from the St Louis Federal Reserve). If you look at the chart of that data in isolation you see a dramatic increase, but that shouldn’t be at all surprising per se: nominal GDP was also increasing and you would expect the amount of money to increase if GDP increases. To examine this properly, I have graphed the US M3 money supply divided by nominal US GDP from 1960:

From 1960 to the mid-nineties, M3 money supply increases at a similar rate to GDP and the ratio between the two oscillates within a narrow range. That ratio begins a ten year downturn after Black Monday in 1987, eventually falling below 0.5x GDP in 1993 and reaching a trough of around 0.45x in the late nineties. China joining the WTO, as described in my blog on the GFC, results in a significant increase in the ratio, but only back to the bottom of the previous range. This was followed by the impact of QE, which can be seen very clearly in the chart. From the beginning of QE by the Fed, the ratio embarks on a steep rise which constitutes the biggest increase ever recorded, taking it through the top of its 1960-1994 range in 2012 and from there into uncharted territory. It has never been this high. Remember that we are comparing M3 money supply to nominal GDP, so any impact of M3 money supply growth on inflation is reflected in nominal GDP in the denominator. No wonder there is too much cash chasing assets right now! For sure, uninvested corporate cash and slow growth is a contributor to the ratio’s rise. But as QE unwinds, normalisation in M3/GDP is likely, despite the excess free cash generated by corporates. That isn’t good news for asset valuations.

Companies can use their uninvested cash to repay debt, which would cause a contraction in the monetary base. The cash returns from those companies to their investors can also be used to repay those investors’ debts or simply allowed to build up in deposit. If those deposits are not lent out, in other words if banks increase their reserves, then, again, the monetary base contracts. Finally, investors can decide to spend the money on goods and services, rather than invest it: the pension fund could increase payments to its pensioners, or the SIPP investor could cash it in and buy a Porsche. Remember QE is premised on a low consumer price inflation environment. If asset prices are high and consumer prices are low, then, to a certain extent, it makes sense to sell assets and consume.

It is therefore unlikely that there will be any “excess cash” chasing investments as QE unwinds. In fact, the reverse is likely. Even if there were, the idea that this would prevent any “buyers strike” is misguided because it ignores investor expectations. Put simply, if there are three investors with cash and one building for them to invest in, they are unlikely to engage in a bidding war if they expect the building to be burned down the following day. Even if cash continues to offer low rates of return and corporate assets offer a higher return than cash, investors will still hold cash if they expect asset prices to fall (i.e. asset returns to rise). In other words, if investors believe returns will rise from 3.9% to 4.4% in six months, it makes sense to wait out, making 0% on cash because you will lock in a 0.5% spread in a year’s time, which will outweigh the lost income for three months in just over two years, with significant enhancements over a longer period (remember our 20 year 50% enhancement from a 1.1% increase in yields, described above).

In addition, what is commonly forgotten is the important demographic component of saving and investing. Given the ageing population and decrease in the proportion of it which is of a natural saving age, asset owners in a post-QE environment cannot expect to benefit from any excess of savings “chasing” their assets, as they have in the past. The youngest baby boomers are now around 54, close to ten years from retirement, with the oldest around 72, currently drawing on their pensions. Assuming most boomers in the developed world will be able to retire at 65, as most working boomers are entitled to, then in three years a majority of boomers will be at a stage in which they are net drawers on rather than contributors to their savings.

 

Risk premium: the clue is in the name

Indeed, it is worth reflecting on the fact that the return premium of corporate assets over government assets, the “risk premium,” should accurately reflect the risks to the cash flows in an economy (the clue is in the name), and questioning whether these risks are as minor as current ultra-low risk premiums imply. Developed market economies have been supported by a multi-decade increase in total debt since the collapse of Bretton Woods, of which the recent dramatic increases post China’s WTO accession are a Gargantuan continuation. Since the GFC and despite any benefit from QE, real median US income has only recently outstripped the pre-crisis level, by around 3%:

Comparing the excruciating, slow recovery in median earnings to the vertigo inducing S&P500 index or the M3/GDP chart above offers a concise demonstration of how QE benefited asset owners rather than the general public.

Whether the economy is able to cope with stable, let alone contracting (if QE is reversed) total debt or not is very much open to question. It is entirely possible that such an environment will make many business models and jobs unsustainable, leading to increased unemployment and bankruptcies. In addition, the concentration of capital in a narrow sector of winners described above has placed increasingly large segments of the economy in a precarious position, most notably retail. In an environment of monetary contraction there is a risk that the troubled sectors decline faster than winners such as Amazon expand. Although, as we outlined above, the positive impact of QE so far has been minimal and, as we argue below, its removal may be healthy for the economy in the long run, I may well be wrong. Or I could be right, but only in the long term, after it has taken a few years for the recovery to take place. A potential impact from the removal of QE is a real risk which should be reflected in the risk premium.

There are also political risks which are worth considering. Global trade, particularly after China’s WTO accession, increased economic activity and demand through debt expansion. While I agree with political attempts to rein in this corporate globalism and feel those who argue it will result in Armageddon are exaggerating, it is very likely to crimp the profitability and returns of large segments of the quoted corporate sector and, indeed, of the narrow portion which has benefited from QE and in which investors have concentrated their portfolios. The high returns earned by the corporate sector, particularly the narrow segment of QE beneficiaries, is also increasingly the target of populist politicians, whether of the Donald Trump or Jeremy Corbyn stamp. While I welcome many aspects of populism and feel that the doom mongering about the impact of such politicians is superficial and self-serving, no one has a crystal ball. The environment for big quoted corporates has rarely been better and the likelihood of a deterioration is increasing. This should be reflected in the risk premium on corporate assets.

 

Should it only be millionaires who can retire?

Finally, it is worth reflecting on the impact of low asset returns on pensioners. An average retiree in the US spends just under $46 thousand a year. US retirees receive payments from social security equivalent to around 40% of average salary. The US median salary is $59,039. This implies that people retiring on the median salary at this level will get ~$30k from social security, leaving them with $16k required from investments to keep up with the $46k average retirement spend. That’s assuming social security pays out (it is currently in deficit in the US and its ability to fund such generous payments is under great threat; the same is broadly true of defined benefit pensions in the UK). To earn $16k from assets yielding 3.9% you need ~$410k or just short of half a million in assets. Crucially, that assumes 100% of assets are invested in equities, one of the highest risk (and therefore highest return) asset classes, which is a very risky approach. Most people’s assets are in their house, which yields far less.

Moreover, this 3.9% return is a cash flow return. The S&P dividend yield is only around 2.1%, meaning that savers would need assets of ~$762k to live from the dividend alone. The dividend is below the cash flow yield because not all of that cash flow is paid out as a dividend; some, indeed not much of it is reinvested in the business (as we discussed above). In theory, that re-invested cash should result in an increase in the share price, allowing retirees to sell some of their shares to maintain themselves without depleting their retirement fund (Warren Buffet regularly points this out in his annual letters). But, as we explained above in painstaking detail, reinvestment in the business only increases the asset value of the company. That increase in asset value is only reflected in the share price if the price/book ratios at which the shares trade do not fall … from their current very high levels. The pensioner’s spending – on food, healthcare, heating, rent – cannot be postponed. It is a regular outlay. If equities de-rate from their current levels that would force the pensioners living off the cash flow yield of their investments to deplete their pension pots in order to maintain their standard of living. Ouch.

The average figures used above only tell part of the story. Houses are so expensive that millenials are increasingly living with their parents. 40% of the US population, apparently, have less than $1,000 in savings (Forbes). How are they going to enjoy anything like a decent standard of living in retirement with assets yielding 3.7% – or 3.9% for that matter?

Some might say that 3.7% growing at 3.2% nominal is a 6.9% return, which is attractive. But there are two issues. First, how much of that growth is real? The latest print for inflation was 2.7%. If that rate of inflation continues, the real return on bonds will only be 0.5% and the real return on equities will only be 4.2%. The real return is equivalent to what you make in retirement after accounting for the increase in your cost of living in retirement. Maybe inflation will be lower than 2.7% in the future. That would be good news for holders of 10 year US treasuries. But it wouldn’t help equity holders, who would see lower cash flows to spend in their retirement.

Second, what is a sustainable level of real growth in an environment of stable or reducing debt? As alluded to above, central banks hope that a resumption of what they consider normal growth will offset any rise in interest rates. However, as we saw above, growth from the seventies to the GFC was supported by a constant increase in debt. It may well be that real growth in economic activity in a stable debt environment remains at much lower levels. However if growth is lower, one might argue, then real yields should go lower, so the net impact on equities would be neutral: the fall in growth would mechanically be offset by the fall in the real discount rate. But for investors, that would create a nominal reduction in the cash generated by their investments and therefore in an impoverishment. That’s what matters to them: the cash they can rely on in retirement. My argument is therefore that the risk premium should be set at a level where investors can earn a decent return, whether the growth rate implied by current US treasury yields is sustainable or not. More importantly, the risk facing equities and other risky assets is a significant rise in risk premiums, on top of any rise in the risk free rate. For these not to result in a fall in share prices, growth in the economy and in corporate profits would have to rise several percentage points. Given the reliance of past growth on increasing debt, this is a very remote prospect indeed.

In basic terms then, having to be half a millionaire to earn a respectable income in old age– and this only by assuming high levels of risk and relying on a social security fund in structural deficit to pay out – does not look like a good deal. Ultimately, the cash flow generated by equities in aggregate has not increased by much in real terms (as outlined in my argument throughout this post that QE withdrawal is more of a risk to asset prices than to the real economy and as illustrated by the real median wage and S&P EPS numbers, above). The increase in equity prices has mostly been due to a compression in the rate of return demanded by investors, under pressure from the flood of liquidity pumped into markets by central banks. But, as pensioners, the cash flow generated by our investments is what we live off. In other words, the rise in asset values has made them unaffordable for those who are not invested without really improving the retirement income of those who have. The strongest argument in favor of asset prices going down and returns for investors going up is that it would allow people to get a decent income in retirement with a reasonable risk – without being fabulously wealthy.

 

A recession-less bear market?

One thing that makes me nervous about predicting such a scenario is that it agrees with a certain pessimistic consensus view, briefly alluded to above. Ten years on from the GFC, it is now common to argue that a repeat is not far away – due to an end to QE:

The right wing Evans-Pritchard argues that debt levels are unsustainable, making a recession and stock market crash inevitable.

He has interesting company in his prediction of doom. Ann Pettifor, a left wing advocate of debt funded government spending, makes a big deal of having predicted the GFC, even though she did so as early as 2003, which is pretty lame: markets were strong for another three years and GDP was above and markets were close to 2003 levels in their post-GFC trough before rebounding strongly.

Anyone who constantly predicts a recession is bound to be right at some point (a stopped watch is right twice a day) – but with huge opportunity cost. She now argues that another crash is on its way, citing similar themes to Evans-Pritchard:

My view is, as described above, that there has been very little benefit to the economy from QE. Apart from create asset price inflation, QE has allowed government debt to increase without there being willing buyers. It has also concentrated capital in the hands of a small group of QE winners. What government, property owners and QE winners all have in common is their dominant positions in the economy. QE, in other words, has benefited a small cluster of rent seekers. At the same time, it has made housing unaffordable for a large segment of the economy and made it difficult for non-property owners to gain access to capital. The aggregate effect has been to stifle creative destruction in the economy. According to statistics I have seen, company formations and bankruptcies combined as a percentage of total companies in existence, in other words the new blood entering and dead wood leaving the corporate sector, is at an all time low. Of course, QE has enabled debt to expand and debt allows economies to spend more than the wealth they generate. Reducing debt will therefore have a negative impact on demand. But that is not necessarily a bad thing. Reduced demand in fact forces producers to become more lean and consumers to spend more rationally. The end of QE will no doubt cause some dislocation, but its main effect will be to accelerate the creative destruction which allows economies to grow without the constant ratcheting up of debt. It will also allow more money to be spent on goods and services and less on rent – to property owners, governments or monopolistic tech companies.

 

Calibrating this forecast

It’s very easy to “call” a turn in the market in a blogpost (or, as in the case of Pettifer, an article). If you’re right, you’re a hero. If you’re wrong, you don’t lose any money. It’s therefore important to explicitly calibrate how confident you are in your forecast. If you calibrate a forecast you get wrong as one of which you are super-confident, you are subject to more humiliation, whereas if you hedge your bets about a forecast you get right you earn less kudos.

I have a low level of confidence in this market prediction. Assets can be cheap or expensive for a long time before markets turn. You always need a catalyst. The main catalyst for reducing the wall of money chasing assets and driving their price up is an end to QE. In other words, to get really bearish on asset prices you need to assume QE is withdrawn. Unfortunately, there is no reason, at the moment, pressuring central bankers to do so. And they like QE, so unless they are pressured they won’t reduce it. Sure, it’s increased wealth inequalities and made property unaffordable in some countries, but QE hasn’t, yet, caused any major problems in the real economy.

 

Chickens and canaries

The risk of QE, in my view, is its impact on capital allocation. If there is no market price for an asset because the government can just print money to buy it, then you will get capital mis-allocation. Prices will not reflect marginal demand for capital but the impact of a political buyer. In such an environment, prices will not correctly reflect risk/reward and you will get capital pumped into sectors with poor returns. Indeed, as I write elsewhere, the broad outlines of QE are identical to the causes of all of the major financial crises in the last twenty odd years. Each of these was caused by an expectation that cheap capital would be continually available in particular regions or sectors, leading to terrible capital mis-allocation: Asian (1997) and Russian (1998) crises, dotcom bubble (2000), sub-prime collapse (2008) and PIIGS sovereign bonds (2009 – date).

The current debacle in the UK restaurant chain sector, which expanded massively helped by low cost of capital due to QE, is perhaps a canary in the coalmine in this respect:

It is worth unpacking the extent to which cheap capital allowing the restaurant chain sector to expand at such a dramatic rate disguised major problems in the UK economy. Local government in the UK depends on business rates to pay for services. But fewer and fewer businesses are able to pay those rates, as they are being driven out of business by Amazon, or retrenching due to digital substitution, like the banks [3]. Local government filled that gap in the UK by granting more and more bar and restaurant licenses to chains like Valerie and Jamie Oliver. These, in turn, were funded at low rates by their investors – they were QE winners after all. This allowed them to open more and more new outlets even if the marginal returns on those new openings were low. In fact, the more outlets they opened the higher went their valuation (and, of course, the bigger the bonuses paid to their management). That allowed jobs to be created and business rates to be paid – everyone was happy (especially the founders after they sold their shares for millions to private equity investors). But after a while the market became saturated with increasingly dull offerings – such as the unspeakably mediocre and now possibly bankrupt Byron Burger (“Proper Hamburgers” – how exciting). Yet the market for eating out was not increasing in size – in fact it was decreasing in some areas due to the influence of companies like Uber Eats and Deliveroo. Eventually, cheap capital led the roll out to go too far. There were too many chain restaurants and not enough diners. Outlets were so far from earning even the low cost of capital previously demanded by their investors that many were making outright losses and bleeding cash. Low cost of capital, as always, disguised the underlying problems in the economy at the cost of massive capital misallocation.

The same is true, mutatis mutandis, of another potential canary: loss making Tesla’s bond yields (in passing I would say that 6.6% on Colombian bonds look very attractive to me):

Tesla is employing around 45,000 people and countless others in its supply chain and allowing various governmental and environmental authorities to claim green brownie points only because it has access to cheap equity capital.

Other UK canaries include:

• Equity release, in which finance companies allow people to remortgage their house on an interest only basis in the hope that the finance company can sell the house to pay off the debt when they die (if the people live for too long, which seems to have been something of a trend lately, the finance company is in trouble).
• Local councils investing in commercial property at high prices (due to QE), taking advantage of low government borrowing costs (also due to QE – spot the circularity?), often in shopping centers which are under threat from internet substitution.

Tesla bonds, chain restaurants, equity release, local authority shopping centers … these may all seem like niche investments which are too small to have an impact on the economy as a whole. But remember the same was true of Asian equities and Russian and Greek government bonds, internet stocks and sub-prime loans. My view is that these chickens will come home to roost at some point (if you’ll forgive the mixed avian metaphor). It’s only a matter of time before enough bad investments enabled by cheap QE capital fail for them to have an impact on the wider economy. In the meantime, assets are very expensive. The return on assets at current prices is low, around 3.7% growing at 3.2% nominal, 0.5% real, as described above. That means the opportunity cost of being wrong for a few years is low. In such an environment, I am happy to miss out on some potential upside in the short term while waiting valuations to offer better returns to me as an investor over the long term.

 

Crowding out the government

 

With government bonds, unlike equities and other risky assets, there is – by definition – no risk premium. Therefore, in theory, it should be impossible for government bond investors to go on a buyer’s strike like equity investors can. The government bond investors need to invest their capital somewhere, one could argue. If one of them is on a buyer’s strike then the other will take advantage of that investment opportunity, reasoning that they are getting a fair risk-free return in a low growth rate and therefore, axiomatically, low return environment. Because they are investing in risk-free assets they have no subjective risk premium to play with and therefore can’t possibly get a higher real risk-free yield than the real growth rate of the economy issuing those assets. In theory.

Although this argument is broadly correct, there is a weakness in the “they need to invest their capital somewhere” part. Some investors, like defined benefit pension funds, are forced to invest a large part of their portfolios in government bonds. Others, however, have more discretion and can invest most if not all of their portfolios in other, risky assets. If the rising risk premium makes those other, risky assets more attractive, then these multi-asset investors have an incentive to allocate capital to those risky assets rather than to risk-free – or even go short the latter to go long the former. The buyer’s strike in risky assets would create an extended, juicy risk premium which effectively “sucked” capital away from risk-free assets. This would, mechanically, push up the real return on risk-free assets above the real growth rate of the economy issuing them (which would have the odd side effect of reducing the risk premium for risky assets without reducing their absolute return).

But, if that happens, then you would indeed get risk-free rates going above the real growth rate of the country and debt costs would eventually become unsustainable for governments. In that case, governments would be forced to reduce in size – just as companies reduce capital if they are in a value destroying industry. In other words, a buyer’s strike in risky assets, which extended their risk premium, could eventually have the knock-on effect of forcing up government borrowing costs to a level at which the state had to shrink. The corporate sector would effectively “crowd out” government borrowing. A shrinking state would be the “collateral damage” of higher returns for investors in a post-QE world.

 

Bond vigilantes

As alluded to above, the concept of buyer’s strike, on which this scenario is predicated, is a strange one in the context of the current period of QE. During QE, it would be ridiculous to envisage such a thing, given the fact that central banks are printing money to buy assets. Far from a buyer’s strike, you have a market dominated by a price insensitive buyer that is printing its own money. The knock-on effect of this has been to relentlessly push the price of assets up and their returns down. Anyone going on a buyer’s strike in such an environment faced a massive opportunity cost, bitter regret and even ridicule. It may therefore be very difficult to envisage this concept at the moment, particularly in connection with government bonds, which were the main beneficiaries of QE.

However, this concept does have a real-life precedent, albeit one now forgotten by all but the more – shall we say – seasoned investors in the market. In the late nineties and up to the bursting of the dotcom bubble in 2001, it was still common to hear people talk about “bond vigilantes.” These were active bond investors (the term was neither comic nor oxymoronic in those days) who exercised their judgement as investors to force the yield on bonds up by selling or even shorting them if they perceived a risk that needed to be priced in. These guys were prepared to sit on low return cash if they thought bond prices were too high. No, really, they were! And this in an era where you could get 6% (or more!) on a ten year bond.

Such vigilantes were mostly active in pushing up the inflation component of the bond yield (if they thought rates were not high enough to contain inflation) rather than the real yield itself. However, when governments behaved irresponsibly, the bond vigilantes would also push up real yields in order to price in the risk of excessive issuance – and even the distant prospect of a credit event. The term “vigilante” was apt because those active investors enforced a kind of harsh discipline on governments, just as the historical vigilantes to whom they were compared did on the remote towns of the Wild West which they policed.

Of course, to use the words “government” and “discipline” in the same sentence today sounds fanciful. No wonder few today can remember the existence of the bond vigilantes who, once upon a time, enforced such discipline.

To quantify the obscurity in which the “bond vigilante” has descended, we can look at the frequency with which it has been used as a Google search term, going back to 2004, the earliest year for which Google makes this data available. The following chart is only a relative chart, and therefore does not tell us the absolute number of searches with the term in any given year. However, amazingly, from 2004 to 2007 there were years when “bond vigilantes” was not used once as a search term by anyone in the world using Google. Although the use of the Google search was not as developed in those days, this is an eloquent demonstration of the insouciance which preceded the GFC:

The term unsurprisingly became more popular as a search term in 2008 when bond investors suddenly realised the horrible risk of the toxic real estate CDOs they were holding. The EU sovereign debt crisis sustained that interest until QE quashed it again. The recent spike in interest, as the Federal Reserve began to indicate that rates would rise again, is notable.

However, this relative interest has to be put into context. When compared to the term “quantitative easing,” “bond vigilantes” doesn’t even register – not even in 2008 when you might have expected “bond vigilantes” to start some kind of fight back. In absolute terms, the bond vigilantes have indeed been forgotten. The term has recently come back into fashion, even being used by the FT and M&G Investments, with M&G notably using the term in an article about another potential canary in the investment coalmine: Italian government bonds.

If QE is going to reverse, we may therefore get a double whammy. First, yields will rise, mechanically, to reflect the absence of a forced, price insensitive buyer – but only to levels which approximate to the low growth rates of the current economic environment and which are still relatively supportive of asset valuations. The second whammy only comes when the absence of such a buyer then allows the market to become more driven by investors’ views on the rate of return they expect on their investment, rather than passively accepting lower returns because of lower growth. As we said, that return is subjective, but, for the many reasons listed above, the environment of plentiful money and low risk is likely to end and there are many serious reasons for investors to demand higher returns. This is likely to be something of a shock to asset owners. The money pumped into the market by QE meant that it had to chase assets. Asset owners have therefore become accustomed to such an environment and may take it for granted. This is likely to be a mistake. Although we don’t know where risk premiums will end up, we can be reasonably confident that it will – at last – be the investors who can be choosy in a post QE world. In such an environment, where investors have the whip hand, we can expect their desire for yield to lead to an increase in the risk premium; in other words, we can expect the buyer’s strike described above to push yields up on risky assets. This may in turn enable the re-emergence of those heroes of yore, the bond vigilantes, shaking the dust off their outfits and pinning on their badges, having been awakened from a long sleep. A posse of those bond vigilantes may then push the real yield on government bonds above the equilibrium real GDP growth rate and force governments to shrink in size. Expect things to get messy.

In conclusion, it is very easy to rationalise the market’s validation of current unattractive asset prices and the low returns they offer investors. However, that rationale is not set in stone. All that is required for it to be undermined is a return of active investor decisions regarding their required returns, as long ago epitomised by the now forgotten bond vigilante. The end of QE, if it ever happens, will at the very least make it possible for investors to exercise discretion about the price they pay for assets. Ultimately, this should result in creative destruction in the economy funding more attractive returns for savers. Although bearish on asset prices, this is a fundamentally optimistic post.

 

 

Footnotes

[1] This general rule equally applies to today’s asset light technology and social media companies, despite the fact that they operate without much in the way of machinery or other tangible capital. That’s because they also need to invest in order to achieve their returns, in technology certainly but above all in very expensive advertising and promotion, which is what allows them to gain the dominant scale which makes their business model profitable. Moreover, rather than invest in factories, companies like Google and Facebook typically invest by buying other companies like Youtube and Instagram to cement their position within their ecosystem.

[2] It is only in quoted companies that we can easily observe, in developed markets, an aggregate valuation above book value. The total corporate sector is also composed of many unquoted businesses, typically small and medium sized businesses, for many of which no valuation is observable, and for many others of which a valuation is only observable intermittently (for example in a private fundraising round or a trade sale, which may only take place once every ten years). We therefore do not have a price and therefore an explicit cost of capital for the whole corporate sector. However, the intermittent transactions in the unquoted sector, which form part of the private equity market, do suggest that valuations of unquoted companies tend to follow those of their quoted sector peers but, typically, with a discount. In conclusion, whatever the extent to which an aggregate cost of capital for the corporate sector as a whole – including both quoted and unquoted companies – is discernible, the intermittent data we have on unquoted companies implies that the corporate sector as a whole currently trades above book value – in other words the corporate sector in aggregate enjoys barriers to entry which prevent new entrants from eroding excess returns.

[3] Landlords are forced to pay rates on empty properties, but their payments are reduced by empty property rate relief. More importantly, the level of rates are determined by rent levels. The key benefit of the restaurant chains to councils was their ability to sustain high rent levels which kept rates in the local authority high.