Monday, August 19, 2019

Gold Oped

Now that 30  days have passed, I can post the whole July 18 gold oped in the Wall Street Journal.

Some of President Trump’s potential nominees to the Federal Reserve Board have expressed sympathy for a return to the gold standard. Conventional monetary-policy experts deride the idea—and not wholly without reason. The gold standard won’t work for a 21st-century monetary and financial system. It is possible, however, to emulate its best features without actually restoring the gold standard.

The idea behind the gold standard is simple: The government promises that if you bring in, say, $1,000 in cash, you can trade it for one ounce of gold, and vice versa. By pegging the dollar to something of independent value, it promises to solve the problem of inflation or deflation. And we don’t need central-bank wizards to run things anymore.

Yet the aim of monetary policy isn’t to stabilize the dollar price of gold; it is, rather, to stabilize the prices of all goods and services. The price of gold has varied from $1,000 to $1,800 an ounce in the past 10 years. Had the Fed fixed that price, critics say the price of everything else would have had to vary this much.

Now, it is likely that if the Fed pegged gold, its price relative to other goods likely wouldn’t fluctuate as much. But the problem is real. Under the gold standard, the U.S. experienced much more volatile annual inflation and deflation than it does today. The gold standard didn’t prevent deflation in the Great Depression and previous panics, a current central concern.

The problem would be worse today. What determines the value of gold relative to all goods and services? In the 19th century, gold coins were used for many transactions. People and businesses had to keep an inventory of gold coins in proportion to their expenditures. If the value of gold rose relative to everything else (deflation), people gained an incentive to spend them, and thereby drive up the prices of everything else. If the value of gold fell (inflation), people needed more of it, so they spent less and drove down other prices. This crucial mechanism linked the price of gold to all other prices.

That link is now completely gone. Other than jewelry and some minor industrial uses, there is nothing special about gold, and little linking the price of gold to all other prices. If the Fed pegged the price of gold today, the price of everything else would just wander away. The Fed might just as effectively peg the price of chewing gum. A monetary anchor is a good thing, but the anchor must be tied to the ship. Gold no longer is.

Broader commodity standards face the same problem. Traded commodities are such a small part of the economy that the relative price of commodities can swing widely with little effect on inflation.

Nor was gold as pure as advertised. Governments didn’t back their money and debt fully with gold reserves. So how could they promise always to exchange money for gold? The gold standard was an art, much like fractional reserve banking. Central banks set interest rates as they do today, to manage gold flows. Central bankers’ pronouncements rattled markets as they do today. There were occasional sovereign crises, featuring runs on governments’ gold promises. Governments couldn’t issue more cash when needed. As a result, there were banking crises and cash was seasonally short around harvest time. In response, the Federal Reserve was founded to “furnish an elastic currency,” not primarily to set interest rates.

The gold standard was really a fiscal commitment, not a monetary one. If people demanded more gold from the government than it had in reserve, the government had to raise taxes or cut spending to buy more gold. More often, the government would borrow to get gold, but governments must credibly promise to raise taxes or cut spending to borrow. This fiscal commitment ultimately gave money its value, not the sometimes-empty promise to exchange money for gold. Taxes ultimately back all government money. The gold standard made this fiscal commitment visible and testable.

With this understanding, the U.S. could enact a policy today that emulates the good features of the gold standard. I call it the CPI standard. First, Congress and the Fed would agree that “price stability” in the Fed’s mandate means precisely that, not perpetual 2% inflation. The Fed’s mandate would be to keep the consumer-price index (or a suitably improved index) as close as possible to a stated value.

Second, the CPI target would bind fiscal policy (Congress and the Treasury) as well as monetary policy (the Fed). Inflation would require automatic fiscal tightening and deflation would trigger loosening, just as a gold-standard government trying to defend its currency must tighten fiscally to raise its gold reserves.

Third, the government would emulate the promise to trade gold for notes in modern financial markets. There are many ways to do this, but the simplest is to commit to trade regular debt for inflation-indexed debt at the same price. Under this system, inflation would cost the government money and force a fiscal tightening in the same way gold once did. And vice versa—the system would forestall deflation as well.

Gold-standard advocates offer a cogent critique of current monetary policy, but a return to gold is unfeasible. A stable CPI, immune from both inflation and deflation, backed by the same fiscal commitments that underlay gold, is worth taking seriously.

Mr. Cochrane is a senior fellow at Stanford University’s Hoover Institution


  1. Requires a government that does what it promises to do. Not likely.

  2. Agreed. Although, the argument could be much simpler:

    1. Pegging to gold doesn't really change anything, because the amount of gold you can get for a dollar could just go down by 2% every year, and we'd still have the identical monetary policy as today. The only way to peg like the way you imply would be to specify that 1 USD is now worth 1/2000 OZ of gold, forever.

    2. Remove the Unemployment thing from the Fed Mandate. Everyone knows the easiest way to get people working is by impoverishing them. That's why we have the 2% inflation. Get rid of the dual-mandate, and we have a more benign system.

    3. Have the mandate be 0% inflation and a maximum CPI volatility. If the ostensible goal is "price stability" how come there's no metric ever researched or published on whether the Fed is actually accomplishing that? Rather, say the goal of the Fed is annual CPI volatility below 2%, up or down.

    4. Although the Fed likes to claim that they reduce price volatility, every trader knows you can't reduce sharpe-ratio of your investments below market without the ability to discover alpha or mispricings. Given that the Fed has demonstrated no ability to generate alpha, they could not possibly be reducing volatility either. That means a gold standard wouldn't be any more volatile even if we pegged the dollar to a specific mass of gold. So that doesn't really belong.

    5. The biggest problem with an actual peg, is that the government will definitely vote to violate that when the going gets tough. If it can't pay social security or whatever, the US public will support a devaluation of the USD, as every democracy typically supports inflation. The only way to do it would be via a constitutional amendment, or, even better, using cryptocurrency

  3. Historically, gold standards have been abandoned when the central bank has become insolvent and can no longer afford to buy back, at par, all the money it has issued. In this situation, trying to maintain convertibility at par will result in a run on the bank.

    A central bank on a CPI standard is subject to exactly the same danger of insolvency, and the same danger of bank runs.

  4. My concern with placing an artificial value on gold is that speculation leads to destructive mining practices with no real benefit to offset the cost to our natural environment. It's for that reason, I'm biased against gold standards. Am I correct in my thinking, or are these two concerns completely separate?

  5. But CPI is not a perfect measure. See the substitution effect.

  6. Well, Dr Cochrane is a PhD economist and highly intelligent and a nice guy. I am just an econo-buff.

    But I disagree with this premise: "Yet the aim of monetary policy isn’t to stabilize the dollar price of gold; it is, rather, to stabilize the prices of all goods and services."

    I think the goal of monetary and macroeconomic policy is prosperity, not necessarily strict price stability (however measured, setting up additional questions about quality of life and the environment. Residents of Los Angeles used to breathe air so thick with smog you might wave hello to your neighbors after a rainy day).

    Suppose, given the structural impediments that appear embedded in Western economies, including most prominently property zoning, that an economy gains greater prosperity with mild inflation?

    Conversely, in a nation enjoying population declines, perhaps natural deflation is the "proper" result. We seen net deflation in cities like Sapporo, Japan.

    George Selgin posits that an n àgricultural economy enduring a crop bust, the wrong policy would be to tighten monetary policy to offset the price increase of that bust. I think large parts of the US today are enduring chronic housing busts, in terms of supply, due to property zoning.

    In any event, I wonder if inflation is really the topic anymore. Globally, interest rates and inflation have been falling for 40 years, until now we see negative interest rates and pretty much dead prices in Europe and Japan. The United States may be next.

    I think the topics now should be how to maintain very tight labor markets and economic growth.

  7. Commodity standards are straightforward to define. Define elemental purity, weight, etc.

    The CPI isn't. Its definition is continuously tweaked by the BLS.

    Why not define the dollar as a fraction of NGDP (say 1 / 21.3T of NGDP)? NGDP is much less susceptible to definition biases than the CPI.

  8. Another view would be that the gold standard influenced the inflation rate primarily through 1) the constraints it placed on bank credit expansion along with the consequent influence on interest rates, 2) the fact that base money depended on the supply of gold at a point in time, and 3) international connections whereby gold outflows prompted banks to raise interest rates which led to reduced demand and prices which then stemmed the outflows thanks to favorable terms of trade effects (and vice versa).

    Sorry but I'm not buying the notion that there's a "crucial mechanism" that's now "completely gone" - a gold standard today could easily restore the three mechanisms above, with gold mostly held in the banking system just as it was during the classical gold standard period. (Households and businesses paid for things and especially costly things with bank notes and checks during the classical gold standard period just as they do today - the fact that coins were composed of precious metals activated Gresham's Law, but I'm guessing you'd find either circular logic or unrealistic assumptions in any attempt to model the mechanism that you're describing.)

    That's not to say I'm defending a gold standard – imo fixed exchange rates would cause problems in today's heavily regulated economies. I think you need laissez faire everywhere for the gold standard to have a chance of working smoothly - once countries go off in different directions in the shape and extent of their governments' involvements in their economies, you get situations like Germany vs. Greece within the Eurozone, where fixing the exchange rate becomes a bad idea because it removes the regular devaluations that such a poorly run economy as Greece needs. You also need to let inflation go wherever it needs to go (no inflation or CPI targeting) to keep international gold flows from being one directional.

  9. Hmm, why is targeting a 0% inflation (a fixed value for the CPI) more like the gold standard than targeting 2% inflation? What is the difference? As well, I think there are good arguments why imposing a tax on holding money through positive inflation is efficient: maintaining a currency system is costly so taxing its users makes sense, and much of the currency is held by people who are engaged in illegal activities or are trying to tax evade so making such endeavors costly seems appropriate.

    Regarding Trump, I think it is ironic that while many of his potential nominees favor the gold standard, which indeed imposes fiscal discipline, he has been asking the Fed to maintain a large balance sheet and lower rates, has been running a large budget deficit (once we adjust for cyclicality) even as debt/gdp is the highest since WWII, and is discussing further cuts in payroll and capital gains taxes. As I wrote in my blog post yesterday, Trump seems to be seeking a "helicopter drop" in order to generate a pre-election boom, whose negative consequences may be long-lasting.

  10. If we assume a completely friction free economy with rational actors there isn't really much of a difference between 2% inflation and 0% inflation. I mean, in theory, we could just denominate our contracts, accounts etc.. I mean if we are fully idealizing and people are perfectly natural nothing magic happens at 0% so if we set 0% inflation we'd just get negative interest rates on our accounts. So in a perfectly idealized world there is no benefit to choosing 0% inflation.

    However, the moment we put things like wage stickiness into our model choosing 0% inflation seems like a really really bad idea. I mean if you targeted 0% how do you propose we allow businesses to reduce real wages when profitability drops? Even if in some alternate world it could be eliminated the mere fact that people rationally expect their nominal wages not to drop has real effects.

    1. Yup, theory says you need to let prices and wages adjust, yet when it comes to monetary policy you have to say just the opposite to stay mainstream. It'd be nice to see those who've questioned this get more attention, such as Borio (BIS paper) and Grant (The Forgotten Depression).

    2. If I'm reading the paper right they don't really give any theoretical argument that deflation is a worthwhile risk to take or that regular, expected inflation is a problem. All they say is that "lower prices may boost real income and wealth." But this boost only occurs to the extent one assumes nominal wage stickiness and under such an assumption this temporary windfall to wage workers always comes at the price of inefficient allocation (rent control for your job will discourage efficient reallocation of talent).

      The rest of the argument in that paper is purely empirical but I have several concerns about that.

      First, I don't think the data quite answer the same question we are wrestling with here. In particular, we are asking the counterfactual question: would it be better if monetary policy favored persistent inflation. The paper measures the correlation between observed deflation and economic outcomes. Even if they found a strong positive correlation between economic gains and deflation it doesn't warrant concluding we wouldn't have been even better off. Even in sustained cases the data may simply be revealing a correlation between productivity gains and deflation (both can be sustained over years).

      To the extent that it does address this question by comparing growth rates under the two conditions the growth rate under inflation is higher than that under deflation in all time periods (table 2). (see below about asset price deflation).

      Second, I can't tell if the output figures are population adjusted which seems important.

      Third, it's very difficult to draw conclusions about the impact of deflation when governments and central banks are trying to manipulate it. For instance, are we learning anything about the 'natural' effect of inflation/deflation or merely when governments have pursued inflationary/deflationary policies?

      Finally, the regression seems too complex given the limited amount of data here. That's supported by the fact that they found few strong relationships suggesting one is largely fitting noise and even still they don't seem to have controlled for all common confounders. Ultimately, I just don't see how controlling for all these terms could even be theoretically useful here. Without a clear model about how we expect asset prices, output and consumer prices to be related I have no idea what these correlations suggest.

      With a model we could just use that model to predict the value of regular inflation directly (regressions just as a model test).

    3. I agree the BIS paper has plenty of limitations as to what it does and doesn’t do, but it seems useful as a counter to the simplistic notion that the Great Depression “proved” deflation causes depression. When the FOMC stresses over core inflation being a few tenths of a percent below 2%, it seems to me that the mainstream mindset depends more on that simplistic empirical belief than on a basic sticky-wage model.

      So my brief comment went in a different direction to what you’re saying is the question we’re wrestling with (I see that after rereading your initial comments). If I were to discuss a sticky-wage model and the implications for inflation targeting, I’d point out that overall employee comp certainly has downward flexibility, so a 0% wage growth floor is unrealistic, that other sets of assumptions lead to sticky prices not just wages being suboptimal, etc., etc., but maybe someone else would take you up on that - I wouldn’t be one to go down that path.

      I’ll definitely review your concerns about the BIS paper again, though, the next time I go back to it – I think I agree with most of them, I just look at the objective differently.

  11. Also note that changes in the relative prices of goods would mean that practically speaking 0% inflation wouldn't make things more stable for any particular individual. Everyone consumes their own goods and no one consumes some representative average basket so I don't see any advantage.


Comments are welcome. Keep it short, polite, and on topic.

Thanks to a few abusers I am now moderating comments. I welcome thoughtful disagreement. I will block comments with insulting or abusive language. I'm also blocking totally inane comments. Try to make some sense. I am much more likely to allow critical comments if you have the honesty and courage to use your real name.