It's sure been a difficult time of late for stock. Over the last decade, the stock market has returned a feeble 0.6% vs. 3.9% for Gilts (and 1.6% for corporate bonds) and bonds have now matched or bettered stock returns over more than 30 years!
In light of this, there has been a lot of questioning recently about the relative attractiveness of shares versus bonds. Some suggest that investors should allocate entirely to bonds, not just because bonds are safer, but because they believe bonds will outperform shares over the long run. In other words, if bonds can deliver higher returns with less risk, what's the point of messing around with shares?
Citigroup wrote a piece in 2009 arguing that: "The cult of equities was dead. Long live the cult of the bond". Citi's theory: a half-century of bias of pension funds towards shares was reversing, and, given the lackluster performance of shares, fund managers were instead turning to fixed-income investments for better returns.
We're mindful of Blaise Pascal's trap that "[p]eople almost invariably arrive at their beliefs not on the basis of proof but on the basis of what they find attractive" but, before jumping on the bond bandwagon, we feel that it's worth being cautious for three reasons:
- Fundamental analysis supports the idea that shares should outperform
- The long run (and really long run) evidence still clearly indicates (by miles!) the superiority of shares versus bonds.
- There's grounds for believing that our recent experience with bonds has been highly unusual (even if it has lasted 30 years!)
What's the difference between shares and bonds?
The essential difference between shares (equity) and bonds is that investing in shares is about buying partial ownership in a company, as opposed to bonds which involve making a loan to it. When an investor buys shares, the value will tend to reflect the earnings experience of the firm — good and bad. In contrast, bonds can never earn more than its face value (plus coupons). shares have (theoretically) an unlimited ability for appreciation but, at the same time, greater downside risk (because they are lower down the capital structure in the event of an bankruptcy). Their returns can be decomposed as:
i) Bond return = Current yield + Capital gain
ii) Stock return = Current yield + Earnings growth + Price Earning multiple change
Thinking about risk/return, it seems clear that shares "ought" to produce higher returns in order for the capital markets to function, otherwise, stockholders would not be being paid for the additional risk they take (this is known as the "equity risk premium").
A Flight to Quality Effect
Despite this, bonds massively outperformed shares in 2011 (15.8% vs. -7.1% according to Barclays). However, there are good reasons to believe that the last few years have been atypical - because of the financial crisis. As Warren Buffett wrote in his 2009 annual shareholder letter:
When the financial history of this decade is written, it will surely speak of the Internet bubble of the late 1990s and the housing bubble of the early 2000s. But the U.S. Treasury bond bubble of late 2008 may be regarded as almost equally extraordinary.
The recent popularity of bonds is in part due to concerns about slow economic growth, coupled with a search for investment income amongst a demographically aging population at or near retirement. However, it also reflects four rather unusual (temporary?) factors:
- A flight to quality effect: Worries over the macro-economic situation and the fate of the Euro have driven people to buy bonds issued by stronger governments (such as those from the UK, US, Germany and Japan) due to "a relative scarcity of safe assets". According to the latest Treasury Department data, foreign investors increased their stake in Treasuries to $4.57 trillion in August 2011 from $4.44 trillion at the end of 2010.
- Consumer deleveraging: Prompted by fears of persistently high unemployment, households have increased savings and cut debt. Bloomberg noted that the U.S. savings rate has tripled to 3.6% since 2005 and has averaged 5.1% since the depth of the financial crisis in December 2008, compared with 3.1% for the previous 10 years, according to government data.
- Institutional deveraging: At the same time, banks are trying to rebuild their balance sheets after taking more than $2 trillion in writedowns from the credit crunch. In the US, they have boosted holdings of government-backed securities to $1.68 trillion from $1.62 trillion in December, according to the Fed.
- Stock Market Volatility: A distrust of stock market volatiity following a decade which saw double-digit losses in four separate years.
Inflation, Interest Rates & the Medium Term
Even over the last few decades, it's undeniable that bonds have had a great run. The Barclays Gilt study shows that gilts have beaten shares over the last 20 years by 1.8% (5.9% vs. 4.8%). Likewise, in the US, long-term government bonds have gained 11.5% annually over the last 30 years, compared with a 10.8% increase in the S&P 500. However, again, this is arguably because of very special factors that are unlikely to reoccur, especially the remarkable falls in interest rates & inflation experienced in the Western World over the last 30 years from the more inflationary environmen of the 1970s. High inflation does not hurt shares like it does bonds (as it is typically passed through to stock holders via earnings, whereas it reduces the purchasing power of a bond investor’s future interest payments and principal). This scenario for bonds seems unlikely to repeat itself in the future, given today’s low interest rate environment (discussed below).
What about the Long Term & the really Long Term?
Looking further back still, there are a number of informative studies, all of which strongly favour equities. Ibbotson Associates' famous analysis of U.S. shares and bonds goes back to 1926. It shows that the S&P 500 compounded to end-2010 at an annual rate of 9.9% vs. 5.5% for long-term government bonds, an excess return of 4.4%. A $1,000 U.S. stock investment in 1926 would have ballooned to $3 million by December 2010 vs. $92,000 for an investment in long-term bonds. Investorsfriend has charted this in real terms:
Similarly, a study by Wharton professor Jeremy Siegel, for example, found that after-inflation returns averaged 7.0% over nearly seven decades ending 1870, then 6.6% through 1925 and then 6.9% through 2004. According to work by Arnott and Bernstein, from 1802–2010, U.S. shares generated a 7.9% annual return vs. 5.1% for long-term government bonds (a 2.8% risk premium). Finally, Barclays' work tracks UK asset returns since 1899 and suggests a 3.6% risk premium (shares have returned 4.9% versus 1.3% for gilts).
Of course, it's also arguable that all these data-sets are a poor guide to the future too. 19th century data may be irrelevant since the US was effectively an emerging market back then. Also, both the UK and US markets enjoyed the status of global hegemons during different parts of the twentieth century, which may have positively skewed the equity upside for both markets...
Still, bonds look pricey!
The last time bonds beat stock returns for a sustained period of 30 years in the US - like they just have - was apparently prior to the Civil War! As Bill Larkin of Cabot Money Management points, "the first thing [this] tells you is you're probably at the most expensive bond market in our lifetime". To think about relative valuation, it's worth turning to the Fed model. This argues that the bond and stock markets are in a state of equilibrium, and fairly valued, when the one year forward looking earnings yield equals the 10-year Treasury note yield. While there are a number of issues with this model (e.g. not all earnings are paid out as dividends, earnings can grow, the treasury yield is nominal while earnings are real), it has a certain weight to it. Applying it in the UK context, the FTSE 100’s earnings yield (earnings per share divided by the share price) is 10% versus a 10 year gilt rate of 2.19%, suggesting that the relationship between the two is completely out of kilter. The position in the US is similarly atypical. Indeed, stock dividend yields and earnings are increasingly above bond yields, prompting the CEO of Blackrock to argue that investors should be 100% in equities.
Jeremy Siegel, finance professor at the University of Pennsylvania’s Wharton School in Philadelphia, went so far as to say that
the rally in bonds is a once in a millennium event, but it’s absolutely mathematically impossible for bonds to get any kind of returns like this going forward whereas stock returns can repeat themselves.
While his mathematics may be a little off, it seems a fair point that bond yields are close to negative territory - which is kind of impossible, because it implies that investors would be willing to pay to lend their money to a borrower. Nowadays, there’s just not much scope for long-term bond prices to go up, since those values are inverse with interest rates, which are currently nearly as low as they can be. It seems more likely that moves by central governments for the past few years to hold interest rates down will eventually end as inflation starts to creeps up. Back in 1981, long-term Treasury bonds had yields that were greater than the average annual total return on common shares. That made long bonds a good bet then, albeit a brave one due to the high inflation rate (8.9%) of the time.
Should an investor favour shares or bonds?
The trouble is that you could have reviewed most of the same evidence outlined above five years ago and decided to invest in shares - and that would have been a bad decision. It is certainly true - as Jim Bianco has rightly observed - that:
The generation-long outperformance of bonds over stocks has been the biggest investment theme that everyone has just gotten plain wrong.
Still, we would argue that it's dangerous to focus too much on the experience of the the recent past ("the fallacy of representativeness"), especially if there's good reason to suggest that it was unusual. In the words of hockey legend Wayne Gretzsky: "I skate to where the puck is going to be, not where it has been”.
Despite ups and downs, stock returns over the last 40 years have been virtually in line with the long-term historical average. On the other hand, bond returns have been not only much higher than their historical averages, but also higher than their current yields. Apart from the recent "flight to safety" (which may admittedly persist for years!), the high bond return effect looks to have been driven by higher inflation in the 1970s and a subsequent declining interest rate environment. The notion of mean reversion - and the apparent absence of any explanation that suggest that "this time, it's different" - implies that investors hoping bonds will continue to outperform in the coming years will likely be disappointed. But over what time horizon - it's hard to say!
Finally, it's worth stressing that the shares vs. bonds dilemma, although often debated, is something of a false choice. In all likelihood, one should look at both. A disciplined asset allocation strategy will considers return and risk tradeoffs but also take into account the diversification benefits of shares, bonds and other securities. One rule of thumb often used is to have your age in bonds and the rest in equities in percentage terms - there's a lot more to it than that but that's a subject for another day!