Traveling to different countries can be eye-opening. One of the things that struck me in all three Latin American countries I visited was the deeply ingrained fear of a return to hyperinflation. The ravages of annual inflation rates of 1000% or more were keenly felt by the population, and the poor were hard hit. While this has not turned them into raving monetarists or Chicago-style free marketeers, they are willing to through some short-term hardship as long as it prevents a return to hyperinflation.

In Santiago I was talking to some people about Chilean government bonds and how they're inflation-protected. That is, the yield you get on a Chilean bond is a real yield, corrected for inflation, rather than a nominal yield like most bonds. Chile isn't the only country in the world issuing inflation-protected bonds, but it was one of the first ones to do so. By now, many governments issue inflation-protected bonds. There are the index-linked gilts in the UK, TIPS in the US, OATi's in France and various other flavors in Sweden and (I think) Australia. Issuance is fairly limited as demand has been surprisingly low for them. You'd think that a risk-free return over inflation would snapped up by institutional and retail investors alike, but in practice the uptake has been limited. Apparently the Treasury in the US is looking at increasing issuance of TIPS. (TIPS stands for Treasury Inflation Protected Securities; the story on the grapevine was that Robert Rubin, then Treasury Secretary, had a fit when he heard people were calling them that. "The US Treasury does not give TIPS!" he is said to have exclaimed.)

Chilean inflation-protected bonds however are not denominated in peso, but in a unit of account called the Unidad de Fomento, or UF. The UF's exchange rate to the peso is calculated on a daily basis relative to the peso, which is the currency in Chile. This exchange rate is based on inflation in the last two months, so that the real purchasing power of one UF remains the same, but it fluctuates in value relative to the peso.

What's remarkable is how widespread the use of the UF is. The UF is used not just for government bonds, but also for rents, house prices, long-term contracts, wages, services and big-ticket items. Its use is deeply engrained in the Chilean economy. Its value is calculated on a daily basis and is widely disseminated. All banks and newspapers carry the conversion rate of the UF to the peso. The value of the UF expressed in pesos depends on the inflation rate of the preceding two inflation data points, so you can project forward the value of the UF until the next inflation data point becomes available. That's why the conversion rate can be calculated for some days into the future.

A large part of Chilean contracts and some goods are denominated in UF rather than pesos. This has the effect of indexing that part of the economy to inflation, moving it out of nominal (inflationary) peso space into real (inflation-protected) UF space. But for all this, the UF is just a conversion factor to pesos. All payments are still made in pesos as there are no UF notes or coins. Whenever a payment is due that was agreed upon in UF, the current conversion rate is applied and the payment is transacted in peso. This means that the functions of money are split into a currency (peso) and a unit of account (UF).

The Chilean Central Bank has published a series of papers on indexation and its effects. The two key documents to read are the introduction and Robert Shiller's look at the UF. They key argument against the widespread use of indexation is that it creates persistence of inflation because it is of necessity tied to lagged inflation information rather real-time inflation data. While an indexed unit of account will create inflation protection in high-inflation environments and can be useful in the transition to a low-inflation regime, it is actually harmful in trying to get inflation into the low single digits. Many of the papers refer to exactly this problem and try to quantify the impact. On the other hand, Robert Shiller is a long-time fan of indexation and argues that the problems of inflation inertia can be overcome.

Shiller argues that all countries should create a unit similar to the UF, and accompany that with another unit indexed to wages. I refer you to his paper linked above for the details. In fact, Shiller goes so far as to argue that a complete monetization of the indexed units of account would be feasible, de facto creating inflation-protected currencies. But what good would it do if all prices were denominated in an inflation-protected unit? Would it not simply reintroduce inflation by another means? Shiller argues that since the monetized units of account would at their base still be convertible to the nominal currency, any inflationary effects would correct themselves in an inflation-protected currency. The least sticky prices would rise in UF and therefore in peso too, be picked up by the inflation rate, which leads to an adjustment of the UF/peso rate which then in turn leads to a downward readjustment of the UF price, pushing the inflation back into the peso and peso only. I'm not sure this would actually work, because if all the UF to peso conversions happen transparently, people will stop thinking in terms of nominal pesos. This in turn will allow inflationary expectations to creep into UF prices.

As a behavioral finance guru, Shiller writes a lot about the illusion of money (again I refer to his article for more detail). The illusion of money is that people feel better off if they have more in nominal terms even if there is no improvement in real terms. The canonical example is that people feel better off if their wages increase only at the rate of inflation. The real purchasing power remains the same, but the nominal amount is higher. This also introduces an aymmetry into the flexibility of any real wage adjustments, as people are extremely unwilling to see their nominal wages reduced. There's a floor at zero, while the upside is unbounded.

We're all accustomed to living in a nominal world although we really should be thinking in real terms. This applies even to people who you might expect to know better such as CEOs and CFOs, who you might assume ought to have some training in these matters. But company balance sheets live in nominal space, their P&L's are nominal, their share prices are nominal and their dividends are nominal. We've seen part of the repercussion of this nominal thinking in recent times as nominal GDP growth in the US had fallen to its lowest level since 1962. As a rough approximation, you might expect a company's top line to grow roughly in line with GDP; as the economy expands, so does the potential for revenue. But companies had gotten used to high nominal growth rates in the economy, and the recession of 2001 coupled low GDP growth with low inflation, resulting in low nominal GDP growth of just 1% or so. This crimped the expectations that companies had for their top line as they saw nominal sales growth decline to unprecedented levels (not many of them were CEOs and CFOs in 1962). An even worse situation exists in Japan, were nominal GDP growth has been negative for the better part of five years now.

Changing the mindset from nominal to real is hard to do and seems to come unnaturally. But in Chile at least part of the economy has been indexed, and people can and do deal effectively with an inflation-protected unit of account, even if it is not fully monetized. Ironically, the Chilean authorities are trying to reduce the reliance on indexation because it hampers efforts to reduce inflation even further. A partial nominalization is underway, as the Chilean Central bank explains on page 29. The UF is not being abolished, but the setting of monetary policy is now based on nominal peso interest rates rather than real UF interest rates. The central bank has also started to issue short-dated debt in peso rather than UF. As long as there is an agreed-upon calculation of the UF the people can continue to use it. So at the moment there is no real danger that the UF will be abolished altogether.

The US Treasury already calculates a daily CPI series based on interpolated data, much like the Chilean UF. In principle it could be used to create an American version of the UF to base transactions on. A more interesting variant would be for a private provider to issue an inflation-protected currency that would maintain its purchasing power relative to the CPI basket of goods. Modern methods of payment (credit and debit cards) make the use of such alternatives technically feasible. But this still would not solve a more fundamental problem with the current monetary system: if the central bank gets policy wrong, the economy will suffer. The central bank has the potential to screw things up in a massive way, and an indexed unit of account like the UF won't help either. There is no price discovery mechanism for money and we have to hope central bankers get it right.

But even if we were to move to a world of real prices by monetizing an inflation-indexed unit of account like the UF in Chile, then that still would not mean an end to price fluctuations. The laws of supply and demand still apply, and if there's a dramatic increase in the demand for milk, then milk prices will rise, UF or no. Only those who buy the exact basket of goods used in the CPI calculation will be fully protected against inflation. And even that protection is imperfect as the inflation data is lagged.

An inflation-indexed unit of account like the UF can be useful as the Chilean experience shows. The current efforts to nominalize the economy are limited in scope and the UF will continue to be used for the foreseeable future. There's also the fundamental question whether it matters if the central bank can get inflation down to 3% from 5% if a large part of the economy is inflation-indexed anyway. The Chilean Central Bank does seem to think it matters.

Chile has become an interesting economic laboratory over the last 20 years, with fully funded pensions and innovations like the UF. Its economy has done reasonably well with far better macroeconomic stability than the rest of Latin America. Thus far the experiments seem to be working. But more importantly, they provide crucial empirical data for economists to ponder.

Now if only I could get a Pisco Sour here...