Volare, oh, oh! (II)
I must admit that I enjoy when a flight is delayed because that grants me time to spare. Time to read, to revisit an issue that you find interesting, or to write Medium stories like this one. In part I — written between December 22 and 23, 2019, while flying from Mexico City to Monterrey — I presented the four measures of volatility studied in the paper by Christos Floros (which you can consult in the references below). I introduced the concept of volatility as a tool to analyze the impact of politics in the economy, by capturing the immediate reactions in financial markets.
In part II (this article), I present — and try to justify — the choice of data, as well as an exploration of the volatility measures for IPC (the main stock index in the Mexican market). Remember that the Structural Break analysis comes in part III.
Data
Remember: this is a conversation about the effects of the executive order to cancel NAIM’s construction, which was announced in late October 2018. The affected sectors are — I assume — the Mexican stock market as a whole, as well as the construction sector in particular. If my guess is true, then we will notice a higher volatility in the indices that track the whole BMV and the stock prices of the construction companies involved in the NAIM project. Of course, other firms may have suffered the consequences of the whole Mexican financial system being under pressure, but — and this is my guess a priori — it would be unimportant. Therefore, in part II we will be analyzing data on the following stocks:
BMV main index:
Stocks of Construction firms:
Stocks of Airport firms:
GRUPO AEROPORTUARIO DEL SURESTE (ASURB)
GRUPO AEROPORTUARIO DEL PACIFICO (GAPB)
Stocks that are not necessarily related to the construction of NAIM:
The idea is to find whether construction and airport firms’ reaction to the executive order was more pronounced than that of the market (overall) and that of other large but [perhaps] unrelated firms. I believe that such a comparison will allow us to distinguish what is NAIM-related volatility from the volatility that is related to other phenomena. In other words, what I am trying to achieve is to isolate the effects of the Presidents’ executive order from that of other events that could have happened at the same time.
Volatility in the Mexican Stock Exchange (BMV)
For starters, let’s take a look at how BMV has behaved and reacted to some of the World’s biggest news of the past 5 years. By all four measures proposed by Floros, we identify the same spikes, the most relevant are:
· August 24, 2015: ‘Black Monday’ in the Chinese stock markets;
· June 24, 2016: Great Britain voted to leave the European Union;
· November 08, 2016: Donald Trump receives 306 ‘electoral votes’ as candidate for the US presidential election of 2016;
· Between 2017 and the first half of 2018, the news are filled with stories about the renegotiation of the North American Free-Trade Agreement (NAFTA, now USMC-A or United States-Mexico-Canada Agreement).
· The fourth quarter of 2018 — what this article is all about — when Mr. Lopez-Obrador assumed the presidency of the Mexican Government.
The Simple Volatility Index overestimates volatility. That is something mentioned by Floros, and it sounds natural: the simple volatility takes into account the least information possible. This is what brought me here, after all: I wanted to know why they publish the other data (Open-High-Low-Close).
Notice in the image above that the time before the December 2018 mark is one of great unrest in the BMV. This is the period that I want to study. That year saw the largest federal election in Mexican history and, as the theory of Political Budget Cycles suggests, that meant many changes to the economic life — normality was broken. Besides, it was the second year of Donald Trump’s presidency. Back then, the news were flooded with messages from Mr. Trump, threatening to abandon the North America Free Trade Agreement.
Now, in the chart above it becomes clear that the Parkinson estimates (green) are the highest of the three other measures. It largely corrects the problem of overestimation incurred by the simple index, but I still don’t like it: it considers High and Low prices only. Garman-Klass, on the other hand, is a more efficient index. Together with the Rogers-Satchell, these are the most efficient measures.
According to Floros, Rogers-Satchell corrects for some biases in both Garman-Klass and Parkinson by incurring in a downward bias. The key difference is that R-S does not assume the drift to be zero. This is why we are examining all four measures instead of just one: none of them are perfect estimators of volatility.
At first sight, it becomes obvious that the news of the airport executive decision had an impact in the volatility of the IPC index. The so-called referendum (in which I participated) started October 25, 2018 and ended 4 days later. The dates (late October) coincide with the visually-apparent beginning of unrest in the Mexican stock market. The effect is higher in Cemex and Pinfra than the others. Also, the effect is higher, more easily recognized in the weeks that came after the referendum.
First, the Garman-Klass index shows the effect is hardly noticeable on the IPC index (bottom-right chart), while Cemex and Pinfra suffer it most.
Finally, the Rogers-Satchell index tells the same story. Again: the effect is higher on firms in construction, then on airports. IPC shows little reaction to the referendum.
Remember, we chose the main stocks of which IPC is comprised and decided whether they are related to the NAIM project or not. However, this implies a strong assumption: that stocks of firms in other sectors (for example, FEMSA, a firm not in construction services) are not affected by policy decisions concerning the construction sector.
I think the evidence studied in this article points in another direction. First, the referendum had some effects on other sectors (albeit weaker than in the construction sector); second, that the effect was more evident in the weeks after the referendum. Of course, this was not the only news affecting the performance of financial securities back in the day, so it is plausible that the combination of the referendum and other news caused the reaction of mid-november.
In the third part we will perform a Structural Break analysis on the series of the stocks mentioned above.
This is all for part II and I hope, thus far, the idea is clear.
Stay tuned!
References
Floros, Christos. (2009). Modelling Volatility Using High, Low, Open and Closing Prices: Evidence from Four S&P Indices. International Research Journal of Finance and Economics. Link to Research Gate
The exercise presented in this article is only intended to illustrate one way in which I would analyze the impact of policy decisions on financial markets’ volatility. It is neither an exhaustive nor a serious analysis. I am not criticizing President Lopez-Obrador’s policy, nor am I praising him. The opinions expressed in this article are exclusively my own, and represent nothing other than my own views.
