What I Think About When Thinking About Mexican Rates

Ok stick with me gang--I'm sorry to go deep on Mexico again but I think it is worth a drill down. While there are some super-talented analysts on the sell-side their incentives to not stray too far from the herd. When big moves happen, cognitive biases can blind people to the big picture. Often times you can extract value by getting into the weeds.

As I noted last week, Mexican local rates rose dramatically in reaction to a spike in CPI, one which was driven in part by the depreciation of the peso and 375bps in hikes from Banxico. Today, rates have rallied well off the highs, but are still sticky at levels above late 2016 levels, both in absolute terms and relative to US rates. For its part, MXN has come all the way back thanks to a weak dollar, EM inflows and higher local rates.

My thesis is that these rates can continue to come down because Banxico can, and will, cut the overnight rate aggressively in 2018. I believe markets are underestimating the magnitude of a rate cut in the same way they underestimated the size of the hiking cycle.  

First is this chart. There is an output gap, and one that I don’t think is going to be soaked up by tepid US demand in the manufacturing sector (note the channel cramming in the auto industry), or domestic demand when interest rates are high and there is uncertainty around the 2018 election.

Similarly, we see that real interest rates are high--Banxico hiked aggressively to stabilize the peso during the EM selloff and Trump-mania.

A more “neutral” real rate, as much as I hate that term, is likely to be around that 2% level--that is still high by historical standards and stands as a relatively conservative estimate given the 5y5y real rate in the US is still stable well below 1%.

This chart also shows that breakevens have been falling consistently since the US election as the “Trump-flation” trade unwinds. While Mex breakevens have been falling too, they are still at elevated levels. As I noted last week I think there is at least 25bps in value in nominal curve built into the breakeven rate (roughly from 3.75% to 3.5%).

What makes me so sure that breakevens shouldn't be permanently higher given headline CPI is currently above 6%? Among other things, wages haven’t shown any sign of secondary impact from that inflation shock--in fact, quite the opposite. This argues that 1) there is slack in the labor market, and 2) businesses are recognizing this spike in inflation as transitory and underlying inflation probably isn’t too far from Banxico’s 3% target.

Next is a similar chart from Banxico for inflation. They expect inflation to top out this quarter and come down relatively fast until it reaches the 3% inflation target in late 2018.

I agree with their path here, in fact I think it could come in on the low side. Let’s look under the hood of the Mexico CPI figures.

A big portion of the increase in inflation came from a hike in gas prices back in January. This was big enough to bring people in the streets in protest. The chart below shows the increase in gas prices alone--not counting second-round effects--has driven headline CPI roughly 1% higher (this is the “incidence”, the increase in the price times the weight in the index). Similar hikes in LP, natural gas, and electricity have had an effect too. Note that all four are near cyclical highs.

Another indirect impact from higher electricity prices is built into the producer price index: industrial, commercial and residential prices. Industrial prices have skyrocketed as the government passes along higher fuel prices and pressures heavy users to increase conservation measures. Just as importantly, commercial electric rates were up over 7%. More on that later.

Three big stories here:

1. After a history of gas subsidies, gas prices will be set by the market throughout the country by the end of the year, and international companies are now free to open service stations to compete with state-owned Pemex. There are some new independent stations that are already selling gas cheaper than Pemex stations, and several international service chains have big plans to expand. This effect combined with lower oil prices could result in 5-10% lower gas prices nationwide.  
2. The electric industry is in the process of deregulation. Outside capital is coming into the market, and the industry in is in the process of switching from dirty, expensive oil, to cheap, clean natural gas. The recent spike in electricity prices is likely to swing negative in the near future.
3. Natural gas is also undergoing a transformation. Thousands of miles of pipelines are going live this year and next to import cheap US natural gas. Again, the price increases here are unlikely to be structural.  

Government controlled tariffs are another big element of CPI. As the name implies, these are prices that are controlled by the federal government. One of the big ones here is hikes in public transportation prices. This is a total one-off, and as the historical chart implies, is unlikely to be repeated.

Moving on to “core”--Merchandise inflation has contributed over 2% to headline inflation. I see this as a mean-reversion--the still slow economy and MXN coming back to mid-2016 levels will cause this component of CPI to head back towards the historical mean, if not below. Lower fuel prices would be supportive here too.

Similarly, the contribution of services inflation to headline CPI is also near historical highs. This is particularly unusual given the low wage hikes over the last year. No big outliers here but given the strength in MXN I think this is another sector that can mean revert, with downside in areas like education, restaurants, airfares, and tourist packages.

Putting it all together--I’ve noted below what I think are conservative estimates for how much inflation can revert from each sector.

Some of this analysis I owe to a paper written by Federico Kochen and Daniel Samano for Banxico. They estimate between .043% and .073% in “pass through” inflation given a 1% change in the usd/mxn rate. My estimates for a reversion in CPI that “unwinds” the impact from the peso deval of late 2016 and early 2017 is broadly consistent with their figures.

Kochen and Samano also had data to show there was a 7.2bp “pass through” to headline CPI from a 1% change in the producers’ electricity price index. Given the 7.3% YoY increase in commercial electricity prices,  I’ve attributed a 50bp decrease in the headline CPI rate should electricity prices remain flat over the next twelve months--again, given the impact of reforms and lower natgas prices, I think this is a conservative estimate.

There you have it--what I believe is a plausible forecast that would get inflation back to 3% by mid-to-late 2018, and largely due to structural factors, while one-offs are unwinding. I would allow that the 50bps in electricity PPI might double count some of the factors I account for individually, but I estimated conservatively to allow for that, and assumed no indirect second-round impact from flat to lower gasoline prices.

What would 3% CPI in 2018 mean for rates? First breakevens would fall by at least 25bps. More importantly, I think Banxico would cut aggressively, consistent with the charts showing a stronger MXN, high current ex-ante real rate and negative output gap. If we were to assume a 2% real rate combined with 3-3.5% medium term inflation, it would not be out of the question to see the overnight rate cut to 5.5% from the current level of 6.75% (or 7%, if Banxico pulls the trigger on a 25bp hike tomorrow). With the entire TIIE curve gravitating around the 7% level, we might see another 50bp move lower if the market buys into the easing cycle.

The 2-5yr segment is the sweet spot of the curve, specifically the 1y rate, 2yrs fwd, which is hanging around 6.80%, implying the overnight rate will be just a touch below current levels by that time. Should inflation return to a 3-handle, that would imply real rates still near current levels, which is out of step with labor market slack, a negative output gap, and a continued trend of low global interest rates.

Certainly the Fed and the presidential election are risks. The election is particularly tricky, because it is tough to see Banxico cutting aggressively in front of that event in July 2018.  But what I think is the real driver here is the reluctance of investors that got burned over the past year to buy into the idea that rates can come back down, especially with structural reforms bearing fruit, listless growth, and commodity prices grinding lower. We’ve seen it happen already in Brazil and Colombia--next it will be Mexico’s turn to cut.

Shawn
TeamMacroMan2@gmail.com
@EMinflationista