Executive Summary: The surge of excited discussion about a return bout with worldwide inflation invites a review of fundamental supply-side principles on the subject, taking into account experience gathered over the last several years. Marginal adjustments are made in our analytical framework, and we've developed some new ideas in how to think of the inflation phenomenon. The general conclusion, though, remains the same as we have been advising: No runaway spiral of inflation is on the horizon, but several more years of moderate increases in the consumer and wholesale price indices reflecting the tapering off of inflationary impulses generated over the last two decades. With a hypothetical 10-year government bond the object of analysis, and equilibrium in the current G-7 range of exchange rates and $400 price of gold, a compound inflation rate of between 2- and 3 percent over the life of the bond seems likely, a higher rate in the near term and a negligible rate near maturity. The oil price appears to be near monetary equilibrium, the pre-1971 gold/oil ratio apparently reasserted. A Bush administration monetary accord in this range of prices would pull monetary risks out of the interest-rate schedule, bringing a major bond rally without a recession, but such a dollar policy has thusfar been slow in developing. Patience is in order.
The recent surge in the broadly watched inflation indices, especially the back-to-back one percent increases in the Producer Price Index, invites a review of where the United States is on the inflation front and what we can expect from here. It also invites a rethinking of basic principles, where inflation comes from and why it ebbs and flows, its relationship to interest rates and credit, and why national interest rates vary. Our general conclusion is that inflation is not the problem that it was a decade ago, that there is no runaway spiral ahead of us, but that for several more years the tapering off of inflationary impulses generated over the last two decades will continue, perhaps with surprising spikes showing up from time to time.
We've made marginal adjustments in our basic inflation model, taking into account experience gathered in the last several years, but our level of confidence in the theory remains high. In November 1983, I reported in "Volcker's October Deflation" on a dinner conversation I had with the Fed Chairman. As you can see, it is precisely relevant at the present moment, right down to the $400 price of gold:
He also worries far more than he should about things he has no direct control over. The consumer price index, for example, is the worst possible guide to Fed policy, because it represents the cumulative errors made by the monetary authorities over a long period of time, years, with residual influences lasting decades. There is simply no way for Volcker to keep the CPI from rising for another several years, unless he is prepared to force cumulative deflationary errors on the economy -- so that he could average out some rising prices with some falling prices. The cost of this kind of victory over inflation would be exorbitant and really not possible; at some pain threshold the government would always be forced to intervene with inflationary bailouts of some sort. If he could be sure of several years of only four or five percent inflation in the CPI, at $400 or so gold, he couldn't be concerned. It's the fear that bigger numbers would show up quickly that keeps him leaning with a deflationary bias.1
As it happened, Volcker's deflationary bias persisted and in the following 14 months, until February 1985, he did squeeze CPI inflation by averaging rising service prices (half of the CPI) with falling commodity prices, oil coming down most particularly. Had he been prepared to swallow "bigger numbers" from time to time, U.S. economic growth would have been faster in the intervening years and the U.S. budget deficit substantially reduced. Instead, we are back to where we were in 1983 after temporarily buying lower CPI numbers that we now have to experience as higher CPI numbers.
So, too, with Alan Greenspan today, telling the Senate recently that although he desires "zero" inflation, he can wait "several years" to get there.2 The Fed Chairman could wrestle back the world oil price through another deflationary crunch, inducing a recession in order for us to get a brief glimpse at lower CPI and PPI numbers. On the other hand, if he knew he could reach his expressed goal of reducing the inflation rate to roughly zero in the next five years, despite some intervening increases of four or five percent annually, he would feel more comfortable than he would with bigger numbers occasionally showing up, inviting hysterical talk of runaway inflation. How can he be sure a steady course will get him to his goal?
A Monetary Phenomenon
As always, we start with the assumption that inflation is a monetary phenomenon. If "money," by which we mean a unit of account, remains constant in value, by definition it is not inflating or deflating. Money does not lose its value because workers demand more of it; they must offer more productivity to get money from employers in order to justify the exchange, or the employer goes out of business and the worker gets no money. Money can only be devalued by government. If a government devalues its money by 20% against some stable standard of value, by definition this means the money will buy 20% less than it did formerly. This means a 20% inflation must occur eventually, although not instantly or evenly for every good and service.
How long does it take between a devaluation and an inflation? The determining factor, I believe, is the average maturity of debt and other contracts expressed in the money. In a country like Peru, which has been experiencing hyperinflation, the average maturity of debt expressed in inti, the national currency, is not more than a few hours. All prices rise in a few hours, locally traded goods and services as well as internationally traded goods and services.
In a country such as the United States, though, where the money still has sufficient integrity to serve as a unit of account for longer term debt, a devalued dollar only loses its value instantly against that which it is devalued. If the dollar is devalued against the yen by 20%, it instantly buys 20% fewer yen, but still buys exactly the same number of hours of a worker's labor, because the worker has a contract that has indebted him to the employer at a fixed wage.
Just because the contract is for a year or two or three does not mean the worker will be able to get a 20% increase as the contract matures. Many other contracts bear upon the factors of production in the worker's industry, and they have to mature as well before the process is complete. There are leases and mortgages and bonds that will gradually mature over decades. Utility rates have to grind their way through regulatory red tape. The issuers of bonds could hold their prices down for a time because the devaluation has reduced the real burden of their debt. Every day in Manhattan, for example, there are enterprises closing up because their favorable rents have jumped with expiration of leases written years ago. Workers with home mortgages also get windfalls as their wages begin to inflate but their mortgage payments do not. It's likely that with our hypothetical devaluation of 20%, the average worker's wage adjusts in piecemeal fashion and is among the last "prices" to fully adjust (not, as Keynesian theory insists, the initial cause of "wage push" inflation).
Fifteen years ago, when I was at The Wall Street Journal, I raised the question repeatedly with academic Keynesians on how they could so casually prescribe deliberate devaluation of the dollar. The stock answer at the time was that yes, a 20% devaluation would eventually raise prices by 20%, but that it would take 20 years or so for that to happen, and it would hardly be noticed at such a low compound rate of interest. Insofar as the past history of the dollar was concerned, they were correct. The 69% dollar devaluation of 1934, the first devaluation since the Civil War, was undertaken expressly to lift commodity prices that had tumbled from an index of 100 in 1930 that to 76.2 in 1933, a horrendous deflation that occurred even as the price of gold was pegged at a constant $20.67.3 [It's my belief that it was the fiscal shock of the Smoot-Hawley Tariff Act of 1930 caused prices to fall temporarily, as in a world-wide going out of business sale, just as prices can surge during a wartime demand and recede in the following peacetime.]
The 1934 devaluation to $35 gold was followed by a brief run-up in the broad price indices during the 1930s. The 1930 index was back to 101.1 by 1941. But after 50 years at $20.67 gold, the dollar had earned sufficient confidence to routinely support contracts of 30 years and more, and the average maturity of debt was perhaps half that. Thus it was 1947 before the general price level had risen as much as the dollar value of gold, as measured by the broad price indices, the wholesale commodity index up to 171.5 in 1947.
Wartime price increases should not be considered "inflation" unless long-term government bonds lose their value. That is, if bondholders are persuaded that wartime price hikes are temporary and will be offset by postwar price declines when normalcy returns, long-term interest rates will ignore the wartime "inflation." This is what occurred in World War I. Against the same index mentioned above, a prewar 1916 level of 98.9 climbed to 178.7 in 1920 before the slide to 113 in 1921 and 100.0 in 1930. U.S. bondholders clung to the belief that, with the government guaranteeing the dollar at $20.67 gold throughout, purchasing power at bond maturity would be preserved.
It was after World War II that the non-monetary theories of inflation took root. The 1934 devaluation, to $35 gold from $20.67, was long out of mind, and the postwar inflation thus seemed to appear out of nowhere. "Pent-up demand" that arose from World War II demobilization fueled the expansion, economists theorized, and prices climbed because of "demand-pull" or "cost-push inflation." Monetarists argue that the postwar inflation was the result of printing-press money created during the war, the Fed's holdings of Treasury securities having climbed to $23.3 billion in 1948 from $2.2 billion in 1941. My belief is that such monetization could not have occurred if the dollar had not been devalued in 1934. It would have led to a rise in the price of gold, but that had already taken place. In any event, although the price level had receded after World War I to the prewar level, it remained on the higher plateau after WWII and could have fallen only if the Fed deflated to the 1933 gold price by demonetizing debt (e.g., buying back Federal Reserve notes by selling T-bills or gold).
In the current context, with inflation again seeming to arise out of nowhere, Keynesians are reviving their cost-push, demand-pull and "wage-price-wage inflation" ideas to explain the recent PPI and CPI advances. True, there is also in the financial press some analysis attributing the PPI increases to the delayed effects of dollar devaluation against the yen, mainly in 1985-86, and there is something to this argument. There is no mention of the price of gold as being a contributing element, the gold price having declined by $100 or so during the last year.
As in the late 1940s, though, we must look deeper into the mechanisms of the world dollar economy to understand the current price movements. Over the last 40 years the dollar has lost more than 90 percent of its value relative to gold and about 80 percent of its value relative to petroleum. At the same time it has lost two thirds of its value against the yen while gaining 20 percent against the Canadian dollar and British pound.
Gold and Black Gold
We can't be sure that gold has the same information properties in this world of floating currencies as it did in the various gold-standard regimes prior to 1971. Still, we can be comfortable with the proposition that it is still the most valuable leading indicator of incipient inflations or deflations, since no other commodity has emerged to challenge gold in that regard. Governments, through their central banks, hold roughly half the world's gold. Petroleum, "black gold," has been the closest in monetary properties. Also a non-renewable, non-perishable, easily-traded resource, more than half of the world's available oil is also inventoried by governments. Gold, we think, is the most sensitive leading indicator of inflation, but it is not a significant component of the producer and consumer price indices. Oil and its byproducts, on the other hand, are a big piece of the PPI and CPI.
Gold and oil got out of synchronization in 1986-87, a period when gold was climbing in price as the Federal Reserve policy of ease joined with Treasury's dollar devaluation moves. The oil price (which had climbed back to $20/bbl after the Fed's 1984-85 squeeze drove gold to $275 and oil to almost $10) had been overtaken by political considerations rather than a pure profit motive. In its effort to bring Iran to a bargaining position in its war with Iraq, Saudi Arabia two years ago allowed supply to swamp demand, thus pushing down the world oil price even as the price of gold remained buoyant. In this period, the dollar had also finally stabilized against the other major currencies in the G-7 accord. As a result, the politically depressed oil price was showing up throughout the industrial world in price indices that lent a false sense of "victory" over inflation.
When the Iran-Iraq war ended a year ago, with the oil price closing in on its 1985 lows and gold about $100 higher than in '85, there was much speculation that oil might drop even further as the two war-torn countries attempted to earn more from oil production to rebuild. Instead, the process by which the oil price has instead climbed back to $20 (thus temporarily reflating the broad price indices) has to be judged the result of the profit motive over the political.
In a completely free market the profit motive overwhelms all other considerations in determining the price that optimizes profits. With myriad producers and consumers, Adam Smith's "invisible hand" arrives at this price at the intersection of supply and demand rapidly, via what he called the "higgledy-piggledy" of the marketplace. In the OPEC cartel, this higgledy-piggledy takes place in slow motion, which students of game theory are most likely to understand. The various members of the cartel have different objectives, but none of these objectives can be completely met to optimize profits for the group -- especially when the time dimension is thrown into the calculus. Iran and Iraq might want to produce more now, to rebuild, but they could easily be punished by Saudi Arabia driving them down the curve of diminishing returns, so that they earn fewer dollars while selling more barrels. So they become reasonable. Saudi Arabia might want to sell more oil now because it has mammoth reserves and a worry that the relative price of oil will continue to decline over time. But if it does so, Iran and Iraq will react by driving the price down the curve of diminishing returns (which looks exactly like the Laffer Curve, by the way). And so it behaves.
In this fashion, we observe that over time cartels, especially government cartels that shift to political goals, are less efficient than the free market in optimizing profit. If this were not true, Mikhail Gorbachev would not now be attempting perestroika in the USSR, replacing glacially slow-motion central planning with more higgledy-piggledy.
All this is important to bear in mind as we contemplate the connections between gold, oil and the dollar. It suggests to us that the price of oil belongs around $20, that it will tend to stay there, if gold stays around $400. We can't be certain that the pre-1971 Bretton Woods oil-to-gold ratio of about 20-to-1 (barrels to ounces) will reassert itself in the 1990s.4 The relative price of oil had been falling over the span of generations and spot Saudi crude was selling at $1.30 in 1970, which would suggest the current oil price might be close to monetary equilibrium. The traditional 16-to-l silver to gold ratio broke down in the 1870s when the United States and France demonetized silver. The old ratio might reassert itself if the world central bankers truly demonetized gold, selling off their huge hoards and possibly sinking gold to $100 (assuming this could be done without undermining confidence), but that isn't on the horizon.
Those who are now arguing that gold has less utility as a monetary commodity because mining technology has improved, and making it cheaper to mine, have to explain the widening gap between the traditional oil and gold ratio, now 20-to-l, down from more than 30-to-l at oil's bottom last year. The tendency, we can at least suggest, is in the direction of only a modestly higher oil price over the next several years. A tightening by the Fed, inviting a general G-7 tightening to fight a myopic view of inflation, would cut global demand for oil via recession. The price of oil would fall, but so would the gold price, and both would rise again as the G-7 central banks eased to lift the western economies out of recession. This general awareness of futility in trying to undo past damage is partly what keeps Alan Greenspan from tightening the screws further.
The Role of G-7 Exchange Stabilization
How much more PPI and CPI inflation do we have to endure if gold is stabilized around $400? This is the same question Paul Volcker asked in 1983. It's likely this question is on Greenspan's mind now. A few things we know for sure: If the G-7 were to formalize cross rates as they now exist, and there were no further monetary disturbances in the seven major currencies, we would have more inflation than Japan and Germany and less inflation than Canada and the U.K. This, for the most part, is why the interest-rate differentials exist between these currencies.
If, for example, the dollar and yen were fixed together precisely by treaty at the current 130 yen per dollar, internal prices would continue to rise faster in the United States than in Japan for some years until both currencies fully absorbed all past inflationary impulses. This is because, as we observed earlier, from a standing start in 1971 the yen price of gold has risen only threefold while it has grown elevenfold in U.S. dollars. Domestic prices in Japan have almost fully offset these inflationary episodes, given the shorter maturity of debt there. It will take longer for wages and prices to adjust here, shifting the purchasing powers of the two currencies back into a more reasonable equilibrium. The best guess we can make about the adjustment process is in the long-term interest rates of the two currencies. There are many elements that comprise a long-term interest rate schedule, but because the dominant component is the purchasing power of the bond at maturity, we can downplay the others in order to arrive at a rough judgment on inflation.
Imagine, for example, the dollar and yen are both equal to an ounce of gold for decades. Interest rates for, say, a 10-year-bond in either of the two currencies would have to be roughly identical. Suddenly, the dollar devalues by 50% and is only worth half an ounce. The dollar interest rate on the 10-year bond would have to rise enough to offset the lost purchasing power of the bond. For this reason, a devaluing country gets no trade advantage, even in the short-run. The increased cost of capital offsets the supposed advantage of devaluation in a lowered price of export goods. Prices of inputs might continue to rise more rapidly in the U.S., yet prices of finished goods remain competitive due to the lower interest expense.
This leads us to posit that if the dollar and yen were both fixed to gold at the current rates, prices of non-traded U.S. goods and services would gradually rise relative to those in Japan, but the cost of capital would gradually decline here, converging with the already low rates in Japan.
For our purposes, the 10-year bond is sufficient, given the recent flatness of the yields in the years thereafter. The U.S. rate of over 9% is currently almost double the Japanese 5% yield. Let's assume a 3% real interest rate, i.e., the average increase that the government bond derives from the productivity of real capital in enterprises (against which it competes). We should not want interest rates to fall below 3% on governments or corporates. This leaves 2% in the Japan coupon to cover the purchasing power component plus all other risk components, and 6% in the U.S.
There are, after all, other risks in holding a bond, the most obvious being the risk of further devaluation of the dollar (or yen) over the life of the bond. A related, separate risk is a deliberately inflationary monetary policy at any time in the next ten years, in an effort to, in effect, repudiate part of the outstanding domestic and foreign debt. Such inflation risk need not imply a smooth trend of 6% inflation, but rather the mere probability of a sudden policy change by some future regime. It is important to emphasize that no amount of current monetary tightening can eliminate this risk of inflation in the distant future. Only a binding, institutionalized commitment to eschew devaluation and inflation as a policy instrument can restore this kind of long-term confidence.5
We're only guessing now, but given this logic, history and recent experience in the bond and commodity markets, we would expect these risks in the 10-year U.S. bond to comprise between three and four percent of the yield. That is, a credible commitment to formalize the G-7 exchange-rate range, anchored by gold in this $400 range, would produce a yield of between 5 and 6 percent. This would imply a PPI inflation rate of between 2 and 3 percent, compounded over the life of the bond. Of course, as we advised Paul Volcker in 1983, we would not expect to see a smooth path, but would imagine the current 4 or 5 percent CPI inflation range sliding gradually to a negligible range in the years approaching maturity of the bond. The only way to squeeze this out would be to deflate the gold price, and as we've suggested earlier, this produces too much unemployment and bankruptcy to be a viable long-term strategy.
In this same analytical framework, we'd make similar forecasts of the other G-7 currencies. The U.K., in other words, can't make much headway in lowering the 10 percent yield on its bonds simply by running budget surpluses. Past devaluations have built more long-term CPI inflation into sterling, some of which has been showing up in recent months with unexpected spikes in wholesale and consumer prices, as in the U.S. Nor is it advisable for the Canadian central bank to avoid a milder run of CPI numbers by periodic episodes of severe tightening.
The advantages to the Bush Administration in reaching an agreement within the G-7 are largely within the realm of debt finance. A monetary accord built around a dollar policy, which is what Jim Baker was moving toward when he was interrupted by the 1987 Crash, would pull the monetary risks out of the interest-rate schedule. We would experience a bond market rally without a deflationary recession. All "dollar debt" problems would be relieved as well, including the S&Ls and Third World.
We're less optimistic that such a plan of action soon will take place in the Administration than we were earlier this year. Jim Baker has not focused at all on international economic policy at State. Nicholas Brady, who is trying to get American businessmen to think long term, is himself snarled in short-run problems of the budget deficit, S&Ls, and Third World debt relief. So is Richard Darman at OMB, who I had thought for sure would have pulled this issue off the back burner by now, being most sympathetic to the arguments made in this paper.
It would not take very much to spark such interest, however. As other options in solving these basic economic problems seem to run out, international monetary policy will get more attention. Questions of the revival of worldwide inflation are also pushing in this direction. The economic summit meeting in June is a likely place for President Bush and his basic economic team to focus on these matters, finding the next step toward reform with the Europeans and Japanese.
Our basic outlook thus remains upbeat. There's no inflation spiral on the horizon. There's a far greater wariness to overreacting to short-term surprises in the price indices than was the case in 1982 or 1984. The major finance ministries, central banks and heads of state seem to be thinking along the same general lines regarding economic growth, exchange rates and commodity prices. We just have to be more patient.
* * * * *
1 Wanniski, Jude, "Volcker's October Deflation," The Political Economy in Perspective, Morristown, N.J. Polyconomics, Inc., November 2, 1983, p. 5.
2 Kilborn, Peter T., "Greenspan Willing to Wait for Years for "Zero" Inflation," The New York Times, March 18, 1989, p. D16.
3 Roy W. Jastram, The Golden Constant: The English and American Experience, 1560-1976, New York: John Wiley & Sons, 1977, p. 146.
4 From 1960 to 1970, Saudi crude averaged $1.36 a barrel, so that gold-oil ratio was 26-to 1. Saudi crude at $16 today is almost exactly at that ratio. In the same period, Venezuelan crude was constant at $1.73, a 20-to-l ratio, also about where it is today. IMF, International Financial Statitistics Yearbook.
5 There is also the risk of outright default -- not so much on government securities, but on high-risk private obligations that compete with government securities. The current tax code, by double-taxing both dividends and retained earnings (through capital-gains tax) accentuates default risk on bonds by making debt an attractive substitute for equity capital. Actual and implied government insurance against losses on debts, whether on LDC debt or the CDs of thrifts, further infects and erodes the government's credit rating as well. Finally, bond markets have to cope with volatility risk, which arises from unnecessarily erratic gyrations in central bank interest rates. Even if the longer-term risks of inflation and devaluation were minimized, the bond's value at any moment (and therefore its liquidity) can nonetheless be adversely affected by periodic, unpredictable spikes in the fed funds rate.