The election race
has begun and investors look optimistic
The purpose of this
article is to go through the scenarios opened last week after official
candidates for the PASO election were presented. This election will be the main
driver for Argentinean bond returns up to year-end, so it’s good timing to do
some math and introduce different election outcomes. Besides defining the
Congress composition, the more relevant aspect of this election is how strong each
candidate will come out for the 2019 presidential election, with particular attention
on the former president Cristina Kirchner and the incumbent president Mauricio Macri.
Three scenarios could be considered in this sense: i) Cambiemos, Macri’s party, achieves a good election in the Buenos
Aires province and Cristina Kirchner doesn’t get a seat for the Senate. This
scenario will put Macri on a great position for a re-election in 2019; ii)
Cristina Kirchner comes up with a great election, raising her as the main peronist candidate for 2019; iii) Cristina
Kirchner does badly in this election, losing most of her political power, but
at the same time Cambiemos doesn’t stand out either, leaving the door open to a
return of the Peronist party under a different candidate.
Along the article
you’ll find an estimation of current default probabilities implied in sovereign
bond prices, which will be useful to estimate expected bond returns under the
different scenarios in consideration. Once we estimate the expected returns for
each scenario, we can guess the probability that matches expected returns with
current bond prices. Presently, investors are giving around 40% chances to the
first scenario, Macri’s victory, and around 30% to the other two scenarios. The
market looks a bit more optimistic than polls around the election, and giving
expected returns under these probabilities, the Argentina 2026 bond provides
the best weighted average capital gain. However, my view is that Cambiemos will a have a tough election
and it’s not obvious that the division of the Peronist party benefits the
ruling party. In addition, the social mood has been heating up, and this could
hurt Cambiemos chances in the election. I would like to see how the campaign develops before
taking an aggressive position towards the election. As a result, I prefer a
conservative approach with a long position in Arg21, which offers better carry
than Arg19 and similar weighted average capital gains.
Calculating default probabilities
To estimate default
probabilities from bond prices I have used the methodology presented in chapter
23 of the seventh edition of Hull’s book[1].
The methodology basically consists in finding a default probability that matches
current bond prices with the present value of the expected loss from a default
along the bond life. This expected loss is calculated by the difference between
the bond market price and the value of its cash flows discounted by the risk
free rate. By applying this calculation to bonds of different maturities, we
can construct a term-structure of default probabilities, where each point of
the curve represents the chance that a default will occur in that particular period.
I call these points “forward default probabilities”. For simplicity reasons we
use Argentina 2019, Argentina 2021, Argentina 2026, and Argentina 2046, but the
job could be extended to the whole universe of Argentinean bonds. I have chosen
default probabilities instead of spreads because the latter includes not only
default probabilities but also other premiums. In fact, this methodology is
useful to obtain the implied value of different bond characteristics such as liquidity,
legislation, or even to estimate the fair value of bonds in different
currencies. Furthermore, it is a better methodology to set up scenarios because
restrictions can be placed on forward default probabilities, leading to more consistent
scenarios along the curve.
Defining scenarios, calculating expected returns, and assessing
what’s priced in
The Argentinean term-structure
of default probabilities shows a jump between 2019 and 2021, signaling that
investors are concerned about the 2019 presidential election, especially when
some sectors among the opposition have been challenging the legitimacy of this
debt. The forward default probability for 2021, at 4.8%, is more than twice the
one for 2019, 1.8% (see Figure 1). After 2021, the term-structure of default
probabilities is increasing but shows a flatter slope, with implied forward
default probabilities of 5.25% and 6.15 for 2026 and 2046, respectively. A good
mid-term election for Cambiemos will
put Macri on a good path to a re-election in 2019, and this will be reflected
by a fall in default probabilities across the curve but specially in the term
between 2019-2023. In particular, I would expect a fall in the default
probability for 2021 to the levels close to the ones currently observed for
2019. The rationale behind this expectation is that investors will start to
price in a second term in office for Mauricio Macri, and according to current
prices, a probability of default under his administration shouldn’t be above 2%
or 2.5%. This probability would lead to a 4.2% positive return in the Arg21
bond.
Figure 1: Implied
default probabilities in Argentinean bonds
Source: Bolsar, Own estimate. Forward default
probability is the chance of facing a default in that particular period. Data from
June 28th.
Default probabilities in the long-end of the curve
would also fall in my view, but by a smaller amount, in a range of 60pp to 90pp.
A more significant reduction might be observed, but I believe the long-end of
the curve is more highly anchored on expectations of macroeconomic fundamentals
and in a lower degree on politics. Structural changes that put public accounts
in order and lead to a sustainable economic growth would be needed before a
larger reduction in long term forward default probabilities is observed. I
believe that
after the election, if Mauricio Macri wins, investors will start to demand not
only signs but also facts showing a reduction in fiscal deficit, a deceleration
in inflation, and a sustainable economic growth, before validating spreads
below 300bp in the long-end of the curve. In this scenario I expect positive
returns of 6.5% and 14.5% in Arg26 and Arg46, respectively, but the return is
mostly related to the fall in the short-end of the forward default probability
curve, as long-end is expected to remain near current levels.
On the opposite side, good election results for Cristina
Kirchner will bring her back to the political scene, placing her as the main peronist
candidate for 2019. This scenario will not be welcomed by investors, and a
sharp fall in bond prices should be expected. I imagine an increase in default
probabilities to at least an annual rate of 5% to 6%, leading to negative
returns in short term bonds between 1% and 7%, and of more than 10% in long
term bonds. To estimate this scenario I have looked at bond prices in 2013. I
don’t believe this situation will be comparable to those times because we are
far away from the next presidential election, but I use those valuations as a
floor to my estimations. I have also looked at valuations on Brazil’s bonds at
the time of Dilma’s impeachment, to see how investor mood behaves under
political uncertainty in the region. In this scenario, Arg19 will provide good
protection as it reaches maturity within Macri’s presidential term.
Table 1: Expected returns under different election
scenarios
Source: Own estimates.
The third and last scenario is one in which
Cristina Kirchner does badly in this election, losing most of her political
power, but at the same time Cambiemos
doesn’t stand out either, leaving the door open to a return of the Peronist
party under a different candidate. I view this scenario as an intermediate one
between the other two, and I expect a positive response from investors. Having
Cristina Kirchner out of the scene for the 2019 election would be good news for
investors, and the focus will turn to who will become the new leader of the
Peronist party. Under this scenario, I would expect some reduction in default
probabilities, mostly among short term bonds. A relevant reduction in long term
bonds is not expected, as in this scenario the Government will have to keep
dealing with the opposition in Congress, reducing the chances to put the
economy on a growing path and correct current macroeconomic imbalances.
After having estimated
expected returns for the established scenarios, we are ready to guess from
bonds prices the probability that investors are giving to each of them.
According to my calculations investors are giving around 40% chances to the
first scenario, a Macri’s victory, and around 30% to the other two scenarios.
To estimate these probabilities, I solve the following system of equations for
pi:
p1, p2,
and p3 are the chances of each scenario, while E(Py,1),
E(Py,2), E(Py,3) are the expected prices for each bond
under each scenario presented on Table 1, and E(PArg19), E(PArg21),
E(PArg26), E(PArg46), are the expected weighted average
price of each bond. I find that Arg26 provides the best weighted average
capital gain[1],
0.37%, followed by Arg21 and Arg19, 0.26%, whereas Arg46 provides an expected
capital loss of -1%[2].
Strategy: A conservative approach, long Argentina 2021
The market looks a
bit more optimistic than polls around the election. My view is that Cambiemos will a have a tough election,
and it’s not obvious that the division of the Peronist party will benefit them
in the election, like most people think. In addition, the social mood has been heating
up over the last months, and this could hurt Cambiemos chances in the election. Under current implied
probabilities and expected returns, the Argentina 2026 bond provides the best
weighted average capital gain, however I would like to see how the campaign
develops before taking an aggressive position towards the election. We are
facing binary outcomes with a lot space for upside and downside, but for now I
prefer a conservative approach with a long position in Arg21, which offers better
carry than Arg19 and similar weighted average capital gains..