A Brazilian Anomaly

One nice thing about being retired is the freedom to spend an irresponsible amount of time looking into topics that catch my eye, without  feeling guilty about it.  For example, the chart below caught my eye during the week. On the right, it appears that the real yield on Brazil's 10-year is greater than the sum of the real yields on the 10 years of the US, China, Japan, Germany, France, the UK, Italy and Australia, not only individual comparisons, but combined, by a margin of about 2:1.



Source: Citi Equity Research:  QE Isn't Working: An Equity Perspective (Nov. 21, 2012) (h/t: Mebane Faber's Idea Farm).

The disparity seemed so anomalous I wanted to understand the reason for it.  Brazil in many respects has a stronger fiscal position than most nations in the sample.  And unlike Italy and the other PIIGS, it has control over its currency, which we have been told repeatedly, is the root cause of the PIIGS' financial problems.  So why does Brazil pay such a spread?

The report containing the chart implies that the reason is "financial repression", a term for government policies designed to repress financial returns to help highly indebted governments reduce their debts -- for example, the gigantic purchases by the Fed and ECB of their respective sovereigns' debt, keeping interest rates in those nations lower than would otherwise be the case as they run large fiscal deficits.  For a nation like Brazil that has a relatively low funded debt-to-GDP ratio of around 35%, presumably financial repression is not the order of the day that it is in more indebted nations.

While this sounded right, when I spent some time looking into the subject, I realized the chart is somewhat misleading regarding Brazil's debt.  In contrast to the other nations in the chart, Brazil's debt is denominated in multiple currencies.  The chart apparently portrays the real yield of its local currency debt (LCD), which is approximately 400 bps according to the chart, although more recently it's shrunk to about 370 bps (approx. 9.2% nominal yield less approx. 5.5% inflation).   But Brazil also has USD-denominated debt outstanding.  And that debt turns out to have a much lower yield, approx. 2.6%, or about 650 bps less than the nominal yield on the local currency debt.  If we subtract out the rate of depreciation of the USD, using the US inflation rate of around 2% as a proxy, we get a real yield on that paper in the 60-90 bps range, much more in line with the rest of the chart. 

But it did get me wondering: why is there such a disparity between the real yields on Brazil's two types of debt. 

It could well be that there are so many dollars slopping around the world and chasing yield that buyers have driven the USD-denominated paper up to an undeserved premium above the LCD, which is consistent with the "financial repression" explanation.  But a look at economies similar to Brazil, like Turkey and Mexico, gets in the way of that easy explanation.  They have inflation rates similar to Brazil, and their USD-denominated debt gives off a similar yield, but their LCD doesn't pay much of a premium to the local inflation rate at all.  Brazil is anomalous in that regard. Also, if it were just that, you would expect a carry trade to arise to arbitrage the two interest rate environments and reduce the disparity over time.

I thought it might be that Brazil just has a low savings rate, which would mean the government would have to pay up to obtain scarce funds.  They do have a low savings rate, but not any lower than Turkey  where yields are no higher than inflation.

I decided to look at whether there was a structural explanation in Brazil's financial sector, which does have some unusual aspects.  Brazilian credit availability is dominated by the state-owned development bank, BNDES, which holds a majority of all the long-term debt in the country.   Its loans are pegged to a different interest rate, designated the long-term rate, and known by the acronym of TJIP, than the central bank's policy rate, the SELIC.  But that long-term rate actually turns out to be lower than the short term rate (an inversion which is itself anomalous and probably  a topic for another day).  So that doesn't explain it either.

It also can't be that Brazil's central bank is trying harder than the rest of the world to fight inflation because  it's actually cut policy rates by 500 bps in the past year, and just cut reserve requirements on top of that, because the economy was basically stagnant last year.  

It's possible that holders of Brazil's LCD just don't fully trust Brazil's official inflation rate - or, more charitably, fear the inflation rate will be even higher in the future, and are charging a premium to hold LCD, which is a garden variety inflation risk / currency depreciation explanation with even some capital flight maybe going on.  Although with the US, European and Japan;s central banks all monetizing debt like mad, is Brazil the biggest risk in that regard?

I didn't find any explanation completely satisfactory, so maybe it is a combination of them all or there are some other local factors at work. If it were just an anomaly, it would be a great carry trade but somehow I doubt such an opportunity has been missed by the investing world 

Along the research trail, I came upon a pretty interesting website, http://brazilianbubble.com/ , which maintains that Brazil is in the final stages of a credit and housing bubble, a thesis that it marshals a good deal of analysis to support; some readers might find it of interest and possible something to keep an eye on in the next year or two.