Over the past thirty years, inflation in the U.S. has averaged just below 3% per year. For many investors, we fear this extended period of price stability has created a complacency about the impact inflation can have on the returns of different asset classes.
But the events that have unfolded since the credit crisis of 2008 should challenge this attitude. First, the crisis reminded investors that the fat tails in financial markets are large – and the previously unthinkable can quickly turn into reality. Second, the crisis sowed the seeds for the possibility of rising inflation. Central banks have increasingly engaged in unconventional monetary policy, and debt levels among developed market governments have ballooned. Monetization of government debt through inflation could be a logical result.
Further, we believe asset prices are much more sensitive to inflation outcomes relative to expectations than actual inflation levels – i.e., investors can react strongly when outcomes differ from expectations. Historically, inflation regime shifts have occurred with little warning. And once a growth spark ignites the inflation gasoline left everywhere by central banks (most recently the Fed with QE3), it may be too late to hedge the effects of inflation.
Therefore, now may be the time for investors who are concerned about inflationary risks to focus on increasing their exposure to asset classes that tend to provide a positive beta to changes in inflation.
While stocks and bonds have generally performed poorly during periods of high and rising inflation, a number of other asset classes have performed relatively well – including commodities, foreign currencies, gold and TIPS.
Over the past five years, the total amount of U.S. public debt outstanding has risen just over 75%, as public sector deficits have been used to offset weakness in the private sector. Meanwhile, as growth has continued to be lackluster, Treasury yields have continued to decline.
However, the process of issuing more government debt at lower yields cannot continue indefinitely. Ultimately the size of government debt relative to GDP must be brought down. Carmen Reinhardt and Ken Rogoff, in their recent paper titled “Debt Overhangs: Past and Present” showed that economic growth in advanced economies with high levels of debt/GDP, on average, lagged advanced economies with lower levels of debt by 1.2% per year. So if a period of robust growth isn’t in the cards, the remaining options are either to reduce the level of debt/GDP through fiscal austerity – a combination of higher taxes and lower spending – or inflation. Given the lack of compromise on the issues of taxes and spending achieved to date in the U.S., fiscal austerity seems unlikely in the medium term, which leaves inflation surprises or default as the most likely end games.
The U.S. is not unique in this respect. Reinhardt and Rogoff have found that over the past 400 years inflation has been the most common way that governments have dealt with excessively high levels of debt.
While the secular case for higher inflation is clear, some may find the deflationary forces of elevated unemployment, a Chinese hard landing, or a disorderly European break-up to be more immediate threats. Although we agree such near-term factors mean inflation is likely to remain subdued over the next year, we also believe the signs of inflationary pressures are already evident. Core inflation is currently running at a seemingly benign rate of 2.1% year-over-year, compared to an average rate of 2.2% year-over-year over the past 20 years. However, over the past 15 years, each percentage point rise in unemployment has, on average, been associated with a 0.4% drop in inflation. Unemployment is at 8.3%, compared to an average of 5.4% over the 1992-2007 period. Given the current high level of unemployment and the large output gap, perhaps the question should be: Why isn’t inflation lower? The current rate of inflation – in the context of 8.3% unemployment – is evidence that the unconventional policies of the Fed are having an impact on inflation.
The seeds of inflation have been sown. As economies slowly rebalance and recover, the inflationary pressures already present today will become more obvious. In this context, in order to formulate a view on asset allocation, it is necessary to turn to history to better understand the relationship between inflation and asset returns.
During the second half of 1979, annualized inflation in the U.S. was above 12%. In six months, the Barclays Government Long Bond Index underperformed cash by 11%, while gold outperformed cash by 85%.
While it’s tempting to discard such volatile periods of inflation – after all, the world was different back then – the link between inflation regimes and asset returns has been relatively stable through history, and it can provide a guide to asset allocation decisions. Unfortunately, there’s little consensus on how to define historical inflation regimes. For example, do simple partitions of the level of inflation suffice (low, moderate, high)? What matters most, the change in the rate of inflation, or the level of the rate of inflation? Do inflation surprises – relative to market expectations – matter more than the actual level of inflation? Is it acceptable to aggregate data across time periods, or should we focus on contiguous, and therefore path-consistent, time periods (the ‘70s for example)?
Academics and practitioners have studied extensively the links between inflation regimes and asset returns, and the appendix to this article provides a thorough list of these references. One of the main takeaways, as it relates to asset allocation, is that historically, a dynamic approach to inflation protection would have typically outperformed a static asset allocation.
In a recent PIMCO article titled “Asset Allocation: Does Macro Matter? Part II” authors Page, Pedersen and Guo built an econometric model designed to improve the accuracy and flexibility of scenario analysis by controlling for macroeconomic expectations.
The power of this approach lies in estimating deviations from market consensus, which drive the sensitivities of risk factor returns to macroeconomic surprises. This study confirmed the positive relationship between inflation changes and commodities returns, and the negative relationship between inflation changes and both stocks and nominal bonds.
The negative beta between stocks and inflation is important given the role equities play in most investors’ portfolios, and it is perhaps the most controversial aspect of inflation-sensitive asset allocation.
Figure 1 shows Sharpe ratios for various asset classes and risk factors during periods of high inflation. Following up on the work of Page, Pedersen and Guo, we define high inflation not in absolute terms, but instead on a relative basis compared to market expectations of inflation at the start of the quarter. In other words, we focus on inflation surprises. We define surprises as the difference between actual inflation at the end of the quarter and expectations for inflation at the start of the quarter. We define a period as having “high inflation” if the inflation surprise is in the highest 25% of historical inflation surprises, based on data from Q2 1973 to Q1 2012. (The appendix provides the list of data sources.)
This figure reveals domestic U.S. financial assets – stocks and bonds – perform poorly when inflation is higher than previous expectations, compared to real assets, foreign currencies and TIPS, which all perform quite well.
If one were to recreate the analysis in Figure 1 using the absolute level of inflation as opposed to inflation surprises, the results would be similar, but less pronounced. Indeed, inflation surprises are a more significant driver of asset returns than just the level of inflation, because it is changes in inflation expectations that tend to matter to the security returns. For example, high inflation isn’t the enemy of bonds. If inflation is high, it is likely that interest rates are high and already factored into the price of bonds. Instead, it is the unanticipated move from low inflation to high inflation that is particularly negative for the returns of bonds or stocks.
To illustrate why changes in the rate of inflation matter more than levels, consider the data in Figure 2, which shows Sharpe ratios for the same assets shown in Figure 1 during the Volcker Fed period of the late ‘70s and early ‘80s. While average inflation rates were high during this period, they were declining both in absolute terms and relative to expectations. During this sub period of high inflation (levels) asset returns were strikingly different than during periods of high inflation surprises – almost a mirror image. This data partition reflects the impact of the Volcker Fed’s commitment to reducing inflation and reinforces the notion that the rate of change in the inflation rate matters more than the actual level.
Given the current macro environment, investors face the possibility that low growth and high inflation may coexist. Figure 3 shows Sharpe Ratios during such a stagflationary regime, defined as periods during which GDP growth was in its bottom 25% and inflation was in its top 25%. It turns out that the assets that tend to do well during a stagflationary period are very similar to those which tend to do well during periods of high inflation surprises. TIPS, commodities, gold, and even bonds performed well during historical periods of stagflation. Positive performance for bonds reflects the “flight-to-safety” effect created by negative growth surprises, despite the greater inflationary pressures. Gold’s performance partly reflects a rally in the early ‘70s which was shortly after the U.S. left direct convertability to gold. Currently, gold trades at historically high levels, suggesting that some expectations for future inflation and continued low growth and negative interest rates are already reflected in its valuation.
Do stocks hedge against inflation?
All of the previous exhibits have shown that stocks have been a poor hedge against inflation surprises historically. To be sure, the relationship between equities and inflation is complicated because inflation is often linked to growth concerns. During inflationary periods that are purely monetary phenomena, stocks may indeed provide a hedge. After all, if twice the supply of money is printed, then stock prices should simply double in nominal price. For example, during the 1921-1923 period of hyperinflation in Germany, stocks protected investors against hyperinflation.
However, in addition to the results we have shown, numerous academic studies have documented a negative beta between stocks and inflation. Yet despite a plethora of academic studies and historical evidence, this negative relationship remains somewhat of a puzzle to financial economists. A common question is whether inflation shocks tend to occur when other equity factors are at play. If so, then in order to assert the existence of a negative correlation today investors must be confident the recurrent – but perhaps not universal – negative factors that give rise to the negative beta between inflation and stocks are prevalent in the current market environment.
Historically, the causation appears to have been that poor economic growth, which (all else equal) lowers stock prices and leads to an increase in government spending, which leads to unsustainable debt-to-GDP ratios, which in turn leads to monetization of the debt through inflation.
Taxes may also play a role in the negative beta between stocks and inflation. During the high inflation of the 1970s, corporations were increasingly taxed on their earnings. In general, government deficits lead to higher expected taxes and thereby lower anticipated after-tax earnings – essentially creating a vicious circle of low stock returns, increased government spending, high inflation (monetization), higher taxes, and … lower stock returns.
These two previous explanations focus on growth effects in equity prices, but inflation should also directly affect the valuations of stocks since they are essentially just a discounted stream of future cash flows. Increases in the rate of inflation are highly correlated with increases in the level of interest rates (See Viewpoint Asset Allocation: Does Macro Matter? Part II) and rising interest rates means a higher discount rate is applied to future cash flows. Holding everything else constant, the higher risk premium for inflation embedded in the discount rate leads to lower equity price-to-earnings ratios.
All these explanations confirm that historically, investors worried about inflation perhaps shouldn’t have bought stocks as a hedge. While most studies were performed on U.S. data, international evidence also supports this conclusion. The negative beta of stock returns to inflation was documented in Japan, the U.K. and in large-scale studies involving major stock markets globally.
It has been suggested, however, that over the very long run, stocks do hedge inflation because they preserve purchasing power. Dividends should keep pace with inflation, and over the long run, inflation goes up, and stocks go up. But an article in the Financial Analysts Journal by Jack Wilson and Charles Jones, published in July/August 1987, reported that over a period of several decades from the end of World War II to the early 1980s, stock returns underperformed inflation.
In the same vein, studies of the correlation between stocks and inflation over longer time horizons (two to five years) confirm the negative beta between stocks and inflation. Also, based on econometric techniques that control for the impact of various other variables on stock returns, negative stocks-inflation correlation has been found up to 70 quarters.i In the current environment of low inflation that is likely headed higher, we believe investors should interpret the both-are-going-up argument with skepticism.
A relatively less studied question is whether certain industries within the stock market provide a better hedge to inflation than others. It has been shown that companies that provide or invest in goods and assets linked to inflation – energy, natural resources and real estate stocks – serve as a better inflation hedge than the broad stock market. In general, companies that can pass inflation in costs on to customers (“flow-through”) generally perform better in inflationary periodsii. Tobacco companies, for example, have a very high flow-through coefficient. (This result makes sense, since customers are addicted to the product.) This effect creates an opportunity for active allocation across the equity space. Investors concerned about inflation can utilize active management to overweight those sectors or geographies that are expected to outperform in an inflationary environment rather than simply adhering to index weights.
To be clear, none of this analysis suggests that an investor should not own stocks or that they are not an attractive asset class that can provide long-term returns due to the presence of an equity risk premium. The main takeaway should be that equities historically have not been a good inflation hedge and that the process of moving from low inflation to high inflation will likely be detrimental to the returns of financial assets like equities and bonds, particularly relative to physical assets.
While commodities as a whole typically provide a good hedge for inflation – in fact, commodity prices are often a source of inflation – one might ask whether certain commodities hedge inflation better than others. Commodities that make up a larger share of the inflation (Consumer Price Index) basket tend to show a higher correlation with inflation. In the case of the U.S. CPI, food and energy make up 25% of the CPI basket, so commodities like oil and corn tend to show higher correlation to inflation than commodities like base metals.
Some investors question the benefits of owning commodities as an inflation hedge because of concerns about indexes all rolling on the same set of days or negative roll yields at the front of some commodity curves (“rolling” refers to investors rolling money out of an expiring contract and into another one, and the process can affect returns apart from whether underlying commodity prices are rising or falling). These are certainly valid concerns, and because of these issues we believe active management is particularly important in commodities. Active management can help to avoid paying a liquidity premium by rolling over a broader set of days than the index and it can also be used to select specific contracts on the curve to minimize the drag from negative roll yield. While all of the asset classes discussed in this paper tend to have a positive correlation with changes in inflation, it is commodities that have had the highest beta. In other words, for each percentage point increase in the level of inflation, commodities have on average the biggest gain – hence they provide a levered response to inflation, such that investors can hold a relatively small amount of commodities to hedge a much larger portfolio.
The world was a different place in the 1970s: Elvis was alive, the Pontiac Trans Am was a popular car, people smoked on airplanes, computer enthusiasts used eight-inch floppy disks to store 300k of data, and inflation was running at over 10% per year. Recently, as investors have become aware of the need to better understand the relationship between inflation and asset returns, they have turned to data from the ‘70s. Instead, our historical analysis suggests a way to partition historical data to tease out expected risk factor and asset returns during different inflation epochs.
Our analysis has confirmed that stocks and bonds have historically performed poorly during periods of high and rising inflation (although during significant growth shocks or stagflation periods, bond duration may still provide a “risk off” hedge), while TIPS, commodities, foreign currencies and gold have performed well. In addition, investors should pay special attention to two important factors:
- How realized inflation may differ from consensus expectations: Surprises in inflation tend to drive asset and risk factor returns more than the actual rate of change in CPI.
- Initial conditions and the path of inflation: Evidence from the ‘70s and early ‘80s suggests that, like good and bad cholesterol, not all inflation is created equal. The rate of change, initial valuations and macro policies matter.
In the case of gold, some may question its ability to continue to hedge against inflation given its relatively lofty price today. However, we believe gold’s valuation is not overly stretched, instead it reflects a world of continued financial repression and negative real interest rates both in the U.S., and globally. As such, we believe gold, which is essentially a currency without a printing press, will maintain its historic correlation to changes in inflation.
Currently there is little in the way on inflation pressures with core inflation running in line with its average of the last 20 years. While it is hard to say with certainty when inflation will move higher, we can identify some of the potential catalysts. A commodity supply shock, such as the closure of the Straits of Hormuz or widespread regional unrest in the Middle East, is one near-term catalyst that could move inflation materially higher. Recall that in the inflationary episode of the 1970s, it was the Arab Oil Embargo in 1973 that caused inflation to double from 5% to 10%.
Absent a commodity price shock, the other likely catalyst to move inflation expectations higher is a gradual economic rebalancing and recovery. As the economy recovers, we expect aggregate demand to increase and the level of unemployment to decrease. As this happens, the Fed will be faced with making a tradeoff between the two components of their dual mandate, price stability and full employment. It is in making this tradeoff during the coming economic recovery that we see the catalyst for inflation. The Fed may err on the side of seeking greater employment and a stronger recovery, believing that temporarily higher inflation can be reversed. However, prior inflationary episodes suggest that reversing prior easing and its impact on inflation expectations may not be as easy as just undoing prior accommodative policy.
When we look forward, over the next several years we see a world of below-average growth and above-average inflation as the developed world continues to deleverage. Central banks globally have been engaged in a series of unconventional policy measures and competitive currency devaluation.
Ultimately, these actions have created an inflationary tail risk.
Historical evidence of inflation regimes:
Kim, Chang-Jin. 1993. “Unobserved-Component Time Series Models with Markov-Switching Heteroscedasticity.” Journal of Business & Economic Statistics, vol. 11, no. 3 (July):341–349.
Kumar, Manmohan S., and Tatsuyoshi Okimoto. 2007. “Dynamics of Persistence in International Inflation Rates.” Journal of Money, Credit and Banking, vol. 39, no. 6 (September):1457–1479.
Page, Sebastien, Mark Taborsky, and Niels Pedersen. 2010. “Asset Allocation: Does Macro Matter?”, PIMCO Viewpoints, www.pimco.com.
Negative beta between stocks and inflation:
Fama, Eugene F and G. William Schwert. 1977. “Asset Returns and Inflation.” Journal of Financial Economics, 5, (November): 115-146.
Fuller, Russell J., Glenn H. Petry. 1981. “Inflation, Return on Equity, and Stock Prices.” The Journal of Portfolio Management, vol.7, no. 4 (Summer): 19 – 25
Stulz, Rene M. 1986. “Asset Pricing and Expected Inflation.” The Journal of Finance, vol. 41, no. 1 (March): 209 – 223
Ma, Christopher K., M.E. Ellis. 1989. “Selecting Industries as Inflation Hedges.” The Journal of Portfolio Management, vol. 15, no. 12 (October): 5027 – 5067
Hughes, Michael. 1992. “U.K. Equity-Gilt Study 1918-1991.” The Journal of Investing, vol. 1, no. 2 (Fall): 37 – 43
Asikoglu, Yaman, Metin R. Ercan. 1992. “Inflation Flow-Through and Stock Prices.” The Journal of Portfolio Management, vol. 18, no.3 (Spring): 63 – 38
Marshall, David A. 1992. “Inflation and Asset Returns in Monetary Economy.” The Journal of Finance, vol. 47, no. 4 (September): 1315 – 1342
Weigel, Eric J. 1994. “Dynamic Links Between Risky Asset Returns and Inflation: The Case of Five Countries.” The Journal of Investing, vol. 3, no. 1 (Spring): 41 -51
Erb, Claude B., Campbell R. Harvey, Tadas E. Viskanta. 1995. “Inflation and World Equity Selection.” Financial Analysts Journal, (November/December): 28 – 42
Watkins, David, David Hartzell. 1998. “U.S. REITs and Common Stock as an Inflation Hedge: A Response.” Financial Analysts Journal, vol. 54, no. 1 (January/February): 87 – 88
Bhardwaj, Geetesh, Dean J. Hamilton, John Ameriks. 2011. “Hedging Inflation: The of Expectations.” Vanguard Research, (March): 1 – 9
Feinman, Joshua N. 2005. “Inflation Illusion and the (Mis) Pricing of Assets and Liabilities.” The Journal of Investing, (Summer): 29 – 36
Amenc, Noel, Lionel Martellini, Volcker Ziemann. 2009. “Inflation-Hedging Properties of Real Assets and Implications for Asset-Liability Management Decisions.” The Journal of Portfolio Management, (Summer): 94 – 110
Geske, Robert, and Richard Roll. 1983. “The fiscal and Monetary Linkage Between Stock Returns and Inflation.” Journal of Finance, 38 (March): 1-33.
Wilson, Jack W., Charles P. Jones. (1987). “Common Stock Prices and Inflation: 1857 – 1985.” Financial Analysts Journal, (July/August): 67 – 71
Study on German inflation and equity returns during the 1921-1923 period:
Lee, S.R., D.P. Tang, and K. Matthew Wong. 2000. “Stock Returns During the German Hyperinflation.” The Quarterly Review of Economics and Finance, 40: 375-386
iFor studies of the correlation between stocks and inflation over longer time horizons (two to five years) see, for example: Bardwaj, Hamilton, and Americks (2001) Amenc, Martellini, and Ziemann (2009), and Watkins and Hartzell (1998). For the econometric techniques that control for various other variables’ impact on stock returns, see Amenc, Martellini, and Ziemann (2009).
iiMa and Ellis (1989) find that energy, natural resources, and real estate stocks provide good hedges relative to the rest of the stock market. Asikoglu and Ercan (1992) discuss the concept of “flow-through”: the fraction of inflation that companies can pass on to customers.