Inflation Myth and Reality

Weekly Market Comment Archive

Investment Research & Insight Page

Inflation Myth and Reality John P. Hussman, Ph.D. All rights reserved and actively enforced. Reprint Policy

The past two years have seen an enormous issuance of new government liabilities. During the two years ended September 30, 2009, the amount of U.S Treasury debt held by the public (outside of agencies such as the Social Security Administration and the Federal Reserve) surged by more than 50%, from $5.05 trillion to $7.55 trillion. During that time, the Fed's holdings of U.S. Treasuries actually shrank by about $10 billion, yet the Fed has explosively increased U.S. monetary base from $850 billion to $2.02 trillion, fueled by massive purchases of Fannie Mae and Freddie Mac's mortgage-backed securities. On Christmas eve, the Treasury quietly announced that it would be providing unlimited bailout funding for Fannie and Freddie over the next three years, since the underlying cash flows received by Fannie and Freddie on these mortgages are not sufficient to keep the agencies solvent.

In total, the quantity of U.S. government liabilities forced into the hands of the public has soared by $3.62 trillion - an increase of 61% since the third quarter of 2007. Keep this figure in mind as various pittances are reported to be returned from TARP funds provided to various financial institutions. Likewise, remember that any interest "earned" by the Federal Reserve on the assets it holds is interest that is either implicitly or explicitly paid by the Treasury, and is returned thereto. Of course, this figure will get progressively larger as government revenues fall substantially short of outlays, and can be expected to do so for years to come.

It is in this context that we should consider inflation risks over the coming decade. At present, inflation risks are hardly considered to be problematic by Wall Street. From the standpoint of the next few years, my impression is that this complacency is probably well-founded, but only because we are likely to observe a second wave of credit losses that will create fresh "safe-haven" demand for default-free government liabilities. From a longer-term perspective, however, I believe that inflation will be a major event in the latter part of the coming decade, with the consumer price index roughly doubling over the next ten years. As exchange rates and commodity prices tend to be more forward-looking and less "sticky" than the prices of goods and services, it is likely that these markets will move substantially well before the eventual peak in CPI inflation.

I have not always held such concerns about inflation or commodities. A year-and-a-half ago, I argued that strong disinflationary pressures were likely to emerge, and I expressed a great deal of skepticism about the sustainability of the commodities surge, which had pushed the price of oil to $150 a barrel (see July 7, 2008 The Outlook for Inflation and the Likelihood of $60 Oil). At the time, the rush to commodities was not based on general inflation concerns, but instead on the view that demand from China and India would drive prices ever higher. Frankly, I'm still unconvinced that China will be capable of enjoying sustained economic growth without shooting itself in the foot as a result of its wholly undemocratic leadership and weakly-developed banking industry. It is not typical that free enterprise "growth miracles" thrive for long in economies in the constant grip of bureaucrats (witness Japan about 1988, and the awkward end of that growth miracle).

So although oil was pushing new highs in mid-2008, I argued that "the problem will emerge a few months from now, as a) economic demand softens further, b) planned production hikes actually emerge, and c) weakening price momentum encourages speculators to close long positions instead of rolling them forward. At that point, I expect that net speculative positions will plunge by 10-15% of open interest and we'll see a sudden glut on the market for spot delivery. It should not be surprising if this speculative unwinding takes the price of crude below $60 a barrel by early next year."

What is different now is that, over time, the massive expansion of government liabilities can ultimately do nothing but undermine the value of the U.S. dollar relative to real goods and services. Whether it undermines the U.S. dollar relative to other currencies depends on the relative propensity for other countries to go on a similar binge of fiscal recklessness. As John Mauldin says, "the dollar may be the worst currency in the world, except for all the others." Indeed, over the near term, credit concerns would likely have the effect of buoying the dollar and pressuring the dollar price of commodities thanks to a safe-haven rush for dollars. While oil has rebounded strongly from the mid-$30's, which it hit earlier this year, my impression is that a second wave of credit losses is likely to take a good deal of the recent progress back before we observe much more sustained gains several years out.

Nevertheless, longer term, the relative supply of U.S. government liabilities will become an overriding factor. To say that the value of the dollar is likely to decline relative to other goods and services is another way of saying that the dollar price of goods and services is likely to increase. Ultimately, crude oil above $200 and gold prices above $2000 will likely owe themselves far less to robust demand in China and elsewhere than to sheer deterioration in the purchasing power of the U.S. dollar a decade from now.

As for the stock market, it is desirable to believe that stocks will prove to be a good hedge against inflation, since they are after all a claim on nominal cash flows which grow over time as prices increase. This was certainly the expectation in the early 1970's, when inflation risks were dismissed by investors in the belief that earnings would keep up with general prices.

Unfortunately, the view that inflation pressures are benign runs opposite to historical evidence. Specifically, there is a relatively high correlation between inflation rates and earnings yields. Investors tend to systematically elevate P/E ratios when inflation rates are low and depress P/E ratios when inflation rates are high. Which is not at all to say that this behavior is rational. To the contrary, the high P/E multiples that coincide with low inflation are also associated with unusually poor subsequent nominal and real returns. Conversely, the low multiples that coincide with high inflation tend to be associated with unusually high subsequent nominal and real returns.

The implication is that (barring actual deflation) investors view low inflation as a "feel-good" indicator and drive stock valuations excessively high in response. Conversely, they tend to punish stock valuations so much when inflation is high and variable that stocks actually tend to deliver very good subsequent returns. Transitions between low inflation to high inflation, then, tend to be quite painful for equity investors, while transitions from high inflation to low inflation tend to be unusually pleasant.

Fortunately, we have enough data both domestically and internationally to analyze inflation-prone environments with the expectation of dealing with them effectively from an investment standpoint. To understand the investment environment is to know how to respond. Presently, the greatest uncertainty for us continues to be the dichotomy between typical post-recession market dynamics and the much more difficult environment that we may very well actually be in, if previous credit crises in history are any guide. We share none of Wall Street's confidence that the more difficult possibility should be dismissed, and suspect that it is premature to declare the credit crisis over.

Inflation Myth and Reality

Will Rogers once said "It ain't what people don't know that hurts them. It's what they do know that ain't true." That observation certainly applies to the myths and realities about inflation.

Mainstream economists and Wall Street analysts hold two very specific views about inflation almost without exception. The first is that inflation is caused by monetary policy, and specifically by excessive money creation. That view is often accompanied by a slight modification that inflation can also be caused by excessive economic growth. The second view is that there is a predictable relationship between inflation and unemployment, known as the Phillips Curve, such that high unemployment is associated with low inflation, and low unemployment is associated with high inflation.

These views are either incomplete or inconsistent with actual economic data. But what's more interesting is that they're actually misinterpretations of economic theory. In order to fully understand this, you have to briefly suspend what you know, and carefully walk through a little bit of economic theory and what it implies about reality. Reality, as it turns out, behaves very much as theory suggests ... and not much as Wall Street believes (a few portions below are reprinted from the July 7, 2008 weekly comment).

To understand inflation, it helps to know a little bit about “marginal utility.” The typical way I used to teach my economics undergraduates was to get them thinking about ice cream. The first cone might give you a lot of happiness. But if you eat a second cone, you'll get a little less enjoyment. The third cone might be just slightly enjoyable. You might be indifferent toward the fourth, and are likely to be averse (negative marginal utility) to eating a fifth. So as you increase the availability of a good, the “marginal utility” – the value you place on an additional unit – declines.

The same basic principle holds for the economy as a whole. Suppose that given the economy-wide supply of ice cream, the marginal utility of ice cream is six smiley faces, and the marginal utility of a pencil is two smiley faces. Given that, the price of an ice cream cone, in terms of pencils, will be just the ratio of the marginal utilities, so an ice cream cone will cost you 3 pencils.

Exactly the same holds true for money itself. If you hold a dollar in your wallet, you might be giving up some potential interest earnings, but you're willing to hold it anyway because that dollar of currency provides certain usefulness in terms of making day-to-day transactions and so forth. If that dollar is held as reserves against checking accounts at a bank, that dollar is implicitly providing a certain amount of banking services. So a dollar bill has a certain amount of marginal utility, by virtue of legal factors like reserve requirements on checking accounts, and convenience factors like the ability to buy a nutty sundae with cash at the ice cream truck.

As a result, the prices of various goods and services in the economy, in terms of dollars, will reflect the ratios of marginal utilities between “stuff” and money.

The dollar price of good X is just the marginal utility of X divided by the marginal utility of a dollar.

So how do you get inflation? Simple.

1) Increase the marginal utility of “stuff”: This happens either if the supply of goods and services becomes more scarce, or if the demand for goods and services becomes more eager

2) Reduce the marginal utility of dollars: This happens either if the supply of dollars becomes more abundant, or if the demand to hold dollars becomes weaker.

Consider how this worked during the Great Depression. Output declined enormously, but the reason was that demand collapsed. The marginal utility of goods and services most likely declined during that period even though production itself was down. At the same time, despite a rapid increase in the monetary base during the Depression, people were frantic to convert their bank deposits into currency, so even the monetary growth that occurred wasn't nearly enough. The frantic demand for currency, resulting from credit fears, translated into a major increase in the marginal utility of money.

So what happened to prices during the Great Depression? Think in terms of the marginal utilities: the marginal utility of “stuff” dropped, while the marginal utility of money soared. The result was rapid price deflation.

In short, inflation results from an increase in the marginal utility of goods and services, relative to the marginal utility of money. It can reflect supply constraints, unsatisfied demand, excessive growth of government liabilities, or a reduction in the willingness of people to hold those liabilities.

This forces us to think twice about the idea that economic growth causes inflation. As it happens, faster economic growth is correlated with lower, not higher inflation. Inflation does tend to become a problem in the later stage of economic booms, but not because the economy is growing too fast. Rather, inflation accelerates because the economy begins to hit capacity constraints and is therefore not able to grow fast enough. The marginal utility of goods rises at the same time that speculation in risky assets encourages people to abandon cash balances, so that the marginal utility of government liabilities declines.

The present situation is mixed. Demand has collapsed, so the marginal utility of goods is depressed. At the same time, there is enough residual risk aversion to keep the marginal utility of money elevated. For that reason, inflation has been contained despite a monstrous increase in the quantity of government liabilities. This situation is likely to endure for a while in the likely event that we get further credit difficulties and sustained unemployment. But it will not endure an eventual economic recovery a few years out. At that point, the huge stock of government liabilities will weigh on the marginal utility of money, while recovering demand presses upward on the marginal utility of goods and services. The result will be a very large and probably sustained inflation in the U.S. in the second half of the coming decade.

Fiscal Policy versus Monetary Policy

The importance of fiscal policy in determining inflation is immediately apparent if we think in terms of the full or “general” equilibrium imposed by a government budget constraint.

See, if you're a banana republic and want to run a huge government spending program, you're not likely to go through the etiquette of issuing government bonds or setting a proper marginal tax policy. You'll just print up pieces of paper. Friedman's first dictum that “inflation is always an everywhere a monetary phenomenon” is largely a reflection of a long history across many countries that heavy government spending financed by printing money predictably leads to inflation. In particularly unproductive economies, it leads to hyperinflation.

But what if the government spending is financed by issuing bonds? It's tempting to think that somehow printing money means an increase in spending power, while issuing bonds means that the government is taking something in return for what it spends, but it's important to focus on the general equilibrium. In both cases, regardless of whether government finances its spending by printing money or issuing bonds, the end result is that the government has appropriated some amount of goods and services, and has issued a piece of paper – a government liability – in return, which has to be held by somebody. Moreover, both of those pieces of paper – currency and Treasury securities – compete in the portfolios of individuals as stores of value and means of payment. The values of currency and government securities are not set independently of each other, but in tight competition. That is particularly true today, when bank balances are regularly swept into interest earning vehicles as often as every night.

To the extent that real goods and services are being appropriated by government in return for an increasing supply of paper receipts, whatever the form, aggressive government spending results in a relative scarcity of goods and services outside of government control, and a relative abundance of government liabilities. The marginal utility of goods and services tends to rise, the marginal utility of government liabilities of all types tends to fall, and you get inflation.

This is important, because it means that the primary determinant of inflation is not monetary policy but fiscal policy .

The chart below presents the historical relationship between U.S. CPI inflation and the U.S. monetary base (4-year growth rates)

Contrast the preceding chart with the relationship between CPI inflation and the growth of U.S. government spending.

Milton Friedman is widely known for two phrases, one which is half right, and one which is exact. The half-right dictum is that “inflation is always and everywhere a monetary phenomenon.” It's half right because a government spending expansion, regardless of the form, will tend to raise the marginal utility of goods and services while lowering the marginal utility of government liabilities. It's very true that the major hyperinflations in history have been triggered by currency expansion, but as long as a government appropriates goods and services to itself in return for pieces of paper that compete as stores of value and means of exchange in the portfolios of investors, you'll get inflation.

The completely correct dictum from Milton Friedman is this: “the burden of government is not measured by how much it taxes, but by how much it spends.”

The Real Phillips Curve

Two weeks ago (Timothy Geithner Meets Vladimir Lenin), I made some remarks about the Phillips Curve that are far enough outside the mainstream that they deserve some evidence. I noted "It is commonly argued that we cannot observe inflation with unemployment so high. In my view, this is a misinterpretation of A.W. Phillips (1958) analysis. While the famed “Phillips Curve” was described as a relationship between (nominal) “money” wages and unemployment, the British data Phillips used was from a period when Britain was on the gold standard, and the general price level was extremely stable. Thus, any wage inflation observed by Phillips was actually real wage inflation.

"The Phillips Curve is simply a standard economic argument about relative scarcity. It says that when the labor markets are tight, nominal wages rise faster than the rate of general inflation (i.e. real wages rise), and when unemployment is high, nominal wages rise slower than the rate of general inflation (i.e. real wages fall). As we observed in the 1970's, high unemployment can exist in concert with high rates of inflation. All that happens, in that case, is that wages tend to rise slower than prices. Assuming labor productivity is growing as well, real wages don't keep pace with productivity growth. In any event, unemployment emphatically does not prevent the inflationary consequences of reckless creation of government liabilities."

Quite simply, the view of the Phillips Curve as a relationship between unemployment and general price inflation wasn't even part of Phillips work, nor is it consistent with the data. This is almost frightening, because we continue to send bright-eyed undergraduate economics students out into the world believing something about Phillips' work that, as Adam Smith might say, "was no part of his intention."

The charts below are based on data since 1947. Monthly unemployment rates were sorted from highest to lowest, divided into equal groups, and the average unemployment rate and year-over-year CPI inflation rate were plotted for each group. What we observe in the data is strikingly opposite to the standard (mis)interpretation of the Phillips Curve. Indeed, higher unemployment is generally associated with higher, not lower general price inflation.

Contrast this with what I would assert is the real Phillips curve, which is simply a statement about labor scarcity. It says, in a very unadorned way, that when labor is scarce (low unemployment), the price of labor tends to rise relative to the price of other things (thus we observe real wage inflation). In contrast, when labor is plentiful (high unemployment), the price of labor tends to stagnate relative to the price of other things (real wages stagnate). Since productivity growth tends to be positive over time, it turns out that real wages actually fall on a productivity-adjusted basis when unemployment is high. From this perspective, it is no surprise that real wages fell by 1.6% in 2009, even as reported productivity grew.