Measuring the Impacts of Argentina’s Presidential Election Process in 2023 on the Stock Market Performance Using a Dynamic Event Study Methodology

Abstract

This study measured the individual and conjoint effects of Argentina’s primaries and first- and second-voting presidential election results, as well as their post-election comparative effects, on the stock market performance of its most relevant economic sectors. Within four different estimation methods, the state-space specification outperformed the rest. The findings suggest that investors can under/overreact compared to post-election sectors performance, the public services sector being the exception. Therefore, those investors who anticipated the election results by liquidating positions in companies in the materials sector and investing more in companies in the energy and other industrial sectors achieved a superior performance.

1. Introduction

Event studies have been some of the most relevant econometric designs in the financial literature (MacKinlay 1997). Their different applications in measuring the effects of corporate decision announcements, regulatory changes, and macroeconomic shocks on stockholder returns are of significant interest to stock market investors (El Ghoul et al. 2023). However, event studies on the effect of announcements related to the presidential election process are relatively more abundant in the literature of developed countries such as the United States (Obradović and Tomić 2017). Diverse U.S. stock market studies have shown that political events generate abnormal stock market returns, but their values and signs can change depending on the winning candidate’s political party (Niederhoffer et al. 1970; Riley and Luksetich 1980; Roberts 1990).

There is some mixed evidence that confirms that when either general economic or public policies announced by a presidential candidate are oriented to positively benefit one particular industrial sector through its stock price fundamentals (and therefore better business expected, free net cash flows for the sectorial firms’ investors, or a lower cost of capital, beyond those of market equilibrium), that industrial sector could show positive abnormal stock market returns at the moment the electoral results are announced (Obradović and Tomić 2017).

On the other hand, event studies on the effect of announcements related to the presidential election process on stock market performance are scarce for developing countries, and there has been practically no evidence until now for the case of Argentina. Some exceptions are documented by (Ratnaningsih and Widanaputra 2019) for Indonesia, (Kumara and Fernando 2020) for Sri Lanka, (Chavali et al. 2020) for India, and (Cázares and Luna 2022) for Mexico, among others.

The study of firms’ stock market performance that comprises the most important industrial sectors considered by the S&P Merval stock market index of Argentina can give important information to local and global investors for portfolio diversification decisions that can contribute to better resource allocation in Argentina, which can also help to improve its current poor economic and financial performance. In addition, the case of Argentina is also a close benchmark for other Latin American developing countries, which share similar roots, language, and political and economic conditions.

As such, the main purpose of this article is to measure the individual and conjoint effects of three events of a political nature associated with the 2023 presidential election process in Argentina, which ended in the election of Javier Milei as president, on the stock market performance of the five most relevant industrial sectors in Argentina. In addition, another complementary objective is to evaluate the five Argentine economic sectors’ stock market performance in a post-electoral period after Milei’s election, in order to corroborate whether the abnormal returns obtained after the announcement of the results during the 2023 electoral process retained their sign and statistical significance during a post-election period.

The above is conducted to compare whether or not the stock market performance, once the 2023 electoral process was over, was aligned with the stock market performance shown within an ex-post-election period, with Javier Milei as president in Argentina (in office since 10 December 2023), and thus evaluate whether or not there was an under- or overreaction by investors in Argentina at the end of the electoral process in 2023.

Thus, this study contributes to the financial literature, first, by evaluating the impact that the results announcement of the 2023 electoral process in Argentina had on the stock market performance of the main industrial sectors that make up its S&P Merval stock market index, based on the Arbitrage Pricing Theory model (APT, Ross 1976) and regarding two orthogonal risk factors, and after detecting the best estimation method between four alternative econometrics specification, corroborating the hypotheses considered for the energy and other industrial sectors; second, by evaluating whether or not the stock market performance of these industrial sectors was maintained after the election of Javier Milei as the new president, in order to visualize whether or not this was just an investors’ sub- or overreaction once the election results were known.

The rest of the article is organized as follows. Section 2 presents the literature review. Section 3 shows the event’s political context and the research design. Section 4 presents the data and methodology. In Section 5, the econometric methods applied in the study are discussed. Section 6 presents the main article findings. Section 7 discusses the results, and finally, Section 8 presents the conclusions.

2. Literature Review

Event studies in finance started developing in the 1970s. The first studies, including those by (Foster 1973, 1975) and (Watts 1973, 1978), analyzed the reaction of stockholders’ returns to the accounting information issued by firms. In light of these pioneering studies, diverse themes have been generated in the field of event studies in finance, including the impact on shareholder returns resulting from announcements related to corporate decisions, regulatory changes, and macroeconomic shocks. El Ghoul et al.’s (2023) recent review of event studies examined 690 new publications in journals such as The Journal of Finance (JF), The Journal of Financial Economics (JFE), The Review of Financial Studies (RFS), and The Journal of International Business Studies (JIBS), using the keyword “event study”.

Table 1. General classification of event studies published in JF, JFE, RFS, and JIBS.

Event/Sample	Country Sample	More Than One Country Sample	Total
Company level	496	49	545
Paired company level	38	0	38
Country level	84	32	116
Subtotal	618 (88.4%)	81 (11.6%)	699 (100%)

Own elaboration based on information from El Ghoul et al.’s (2023) study.

presents a summary of their respective classifications.

At the company level, event studies were classified based on the impact on stockholder returns resulting from merger and acquisition announcements, restructurings, litigation, and securities titling. Companies were classified based on financial stress in bank–company relationships, losses, profit announcements, and debates on approaches to using market portfolios as benchmarking. At the country level, they were classified based on government reforms/legislative changes, changes in sovereign debt ratings, and tax reforms. Less than 12% of event studies reported in Table 1 (only 4 (0.57%)) are related to events associated with government elections or political risks.

For the case of developing countries, Ratnaningsih and Widanaputra (2019) sought to determine the reaction of the capital market to the announcement of the 2019 presidential election results in Indonesia based on the companies belonging to the KOMPAS 100 stock market index, which considers any deviation from the average in the estimation period to be abnormal. They concluded that the elections had a relevant impact on this stock market index.

Kumara and Fernando (2020) investigated the impact of political events on the performance of the daily index of the Colombo Stock Exchange in Sri Lanka and how these events affected the returns of the All Share Price Index (ASPI) in response to other events from May 2009 to March 2018. They examined political events using the adjusted average model or the adjusted yield to the average and concluded that investors reacted excessively to good political news and insufficiently otherwise. Nonetheless, the impact direction depended on the event’s constitution and market favorability.

Chavali et al. (2020) examined the power that elections between the years 2014 and 2019 held in the stock market of India, a country with an emerging economy, which is considered one of the largest democracies in the world. They evaluated a sample of 31 companies listed on the Bombay Stock Exchange, analyzing whether the impact on the market would be the same when a party gains power and comes to power for a second period. One of their conclusions showed that the impact on the market is not the same between two elections, despite bringing the same party to power for a second time. The study confirmed that elections positively impact the stock market; however, the stock market’s reaction is more powerful when a party comes to power for the first time than the second time.

Cázares and Luna (2022) analyzed the impact of uncertainty on the Mexican stock market regarding who will be the elected president in that country, taking a sample of three electoral periods (2006, 2012, and 2018). Using a GARCH model for the data analysis, they concluded that electoral uncertainty is a substantial factor that generates changes in financial markets, as both general political ideologies (left and right) generate volatility in the market. The left-side ideology is reflected at low levels in the stock market returns, while the opposite-side ideology is complemented by a wealth investors’ increase in the stock market.

For the U.S. case, Niederhoffer et al. (1970) and Riley and Luksetich (1980) showed that over long-run time series, the contemporaneous stock market reaction mainly depends on the winning candidate’s party. Roberts (1990) concluded that Ronald Reagan’s victory in 1980 created positive abnormal returns on the military sector. Obradović and Tomić (2017) concluded that Donald Trump´s victory in 2016 created positive abnormal returns on the military, financial, and energy sectors. However, statistically significant abnormal returns were not found for the electronic and health sectors.

Previous evidence shows a research gap, motivating the study of the particular case of Argentina and especially its economic sectors. Argentina is a rich country in resources that, before the presidential election of 2023, walked into worsening inflation, poverty, unemployment, and country risk indicators (Spruk 2019), and it is also characterized by a stock market rated below the border by agencies such as MSCI (www.msci.com, accessed on 31 October 2024), with a history of high economic, political, and social instability, therefore making it interesting to evaluate how the different Argentine industrial sectors were affected in terms of their stock market valuation, given the announcement of the results of the 2023 presidential election process, which ended with the election of Javier Milei as president. The evidence of this study will help to answer whether or not strategic sectors, such as the financial and energy sectors, among others, present a superior stock market performance conditional on a single presidential candidate’s political party when the candidate is finally elected. In addition, because this study is pioneering in this field in Latin America, especially regarding its methodology, it will surely be a benchmark for comparative research that could be developed in the future in similar regional developing countries.

We also hypothesize, based on the previous evidence, along with Milei’s economic plan (see further in this article), that while the announcements of the results of the 2023 Argentine electoral process were reaching the Argentine stock market, the semi-strong efficient market hypothesis (due to the economic informational content associated with Milei’s arrival to the presidency of Argentina) should be demonstrated by observing positive or negative abnormal returns depending on the evaluated industrial sector.

We hypothesize positive abnormal returns or a superior stock market performance for the financial sector, given the announcement of strong inflationary control and greater stability in the value of the Argentine peso thanks to the dollarization of the economy, and the same for the energy sector, given the announcement of policies aimed at reducing taxes on companies, a greater transversal incentive for competition and efficiency of the productive sectors, where the energy sector plays a key role as a supplier. We also hypothesize positive abnormal returns in the sector of various industries, since this represents a diversified portfolio of companies mainly concentrated on services (retail and others) that will seek to grow and contribute to create economic value based on the incentives announced by Milei (tax reduction and a search for greater efficiency in the participants of this sector).

Regarding the materials and public services sectors, negative abnormal returns are hypothesized because they are sectors with few participants, not very competitive within the industry, and traditionally protected with subsidies or tariff policies that are convenient for them, which was in line with Massa’s economic proposal but not with Milei’s economic plan.

Fama (1970) established that when the form of efficiency is strong in the stock market, the prices of shares, and, therefore, their returns, must reflect all historical, public, and private information, which makes the presence of abnormal returns difficult.

A possible explanation for the presence of abnormal returns at a given moment is the existence of information pieces, which can take time to incorporate into stock market formation prices. This threatens strong efficiency but not stock market’s semi-strong efficiency, which can be present in the electoral presidential process, as in Argentina in 2023. In this regard, the definition of the model that generates equilibrium returns or normal returns plays a fundamental role, because any market equilibrium deviation can be evidence against the strong form of market efficiency.

The Capital Asset Pricing Model (CAPM; Sharpe 1964; Lintner 1965; Black 1972) and the Arbitrage Pricing Theory (APT; Ross 1976) are the most referenced theoretical models in the literature and have been used empirically to generate stock market equilibrium returns (Ortas et al. 2015; Jorion 1991). In its empirical version (derived from a fair-game model), either CAPM or APT takes a linear specification, which can be modified in order to capture the potential presence of abnormal returns at some point in time and therefore generate evidence against the strong form of market efficiency. Through this article, a linear, modified APT specification was estimated under four different econometric estimation methods that will be presented in the Methodology Section of this article. The reasons why a two-APT-factor specification was used, instead of a one factor market model based on CAPM, are discussed next.

The CAPM, at a theoretical level, assumes investors who can diversify their resources in a market portfolio that is efficient in the risk-expected return space, thus being able to eliminate all non-systematic risks present in the variability of the stock returns of their investments. However, this market portfolio is difficult to observe in practice (Roll 1977), and therefore the results based on the CAPM must be interpreted with caution, since its empirical tests regularly use approximations of the true (non-observable) efficient market portfolio. This problem is more prominent to the extent that stock markets present greater segmentation, as is the case with Argentina, which currently has a low rating (less than emerging), according to specialized financial organizations such as MSCI.

The APT, on the other hand, does not require the assumption of the existence of a market portfolio that is efficient in the risk-expected return space. This indicates that the stock returns of a portfolio may be influenced by a finite, but indefinite, set of macroeconomic risk factors, orthogonal to each other, which prevent arbitrage (Ross 1976).

Since the stock returns of Argentine companies and industrial sectors could be affected, beyond the systematic risk of their stock market (S&P Merval is used as the Argentine stock market proxy in this article), and in order to evaluate the influence that political risk may have on the econometric estimates presented in this research, a linear model with two orthogonal risk factors was established, which also allows incorporating the presence of abnormal returns at the time the results of the presidential electoral process experienced in Argentina in 2023 were known. The orthogonality between the risk factors, market and political risks, is important, based on APT’s fundamentals, in order to avoid redundancy and thus eliminate a duplication in risks.

To construct the orthogonal component of political risk, with respect to market risk, a simple linear regression was run between the variable used as a proxy for political risk (which will be indicated later) and the excess returns of the S&P Merval. The residuals of this regression were then obtained, which by construction captures the orthogonal component of political risk with respect to the Argentine stock market risk, a component that is subsequently used as a second risk factor.

Incorporating political risk is especially relevant in developing countries, since it is a determinant in attracting foreign investment. This can have an impact on the development of the capital market in a country, because its dynamics can be altered as a result of changes in the political and economic systems that occur when a new president arrives (Siregar and Diana 2019). Political risk is defined as the risk of operating or investing in a country in which political decisions can eventually generate significant economic losses for investors. Political risk can, in practice, take several forms. According to the political risk report issued by the WIPR World Bank (2013) and shown in Figure 1, the political risk forms of greatest concern to investors in developing economies are risks associated with adverse regulatory changes, violations of the agreed-upon terms in contracts (breach of contracts), restrictions on the terms and conditions of contracts (T&C restrictions), civil disturbance, non-honoring financial obligations (NHFO), expropriation, terrorism, and finally, war.

Although it is difficult to find a single comprehensive indicator of political risk for each emerging economy (Bekaert et al. 2015), there are specialized financial institutions that publish country risk indicators that focus especially on the political risk of non-compliance with financial obligations contracted by countries, which is one of the political risks that are worrying for investors in developing economies such as Argentina. In this way, a direct relationship between both risks should exist.

One of these institutions is JP Morgan, which prepares a country risk indicator that measures the difference paid by U.S. Treasury bonds against those of the rest of the countries. In this way, the index measures the surcharge that an Argentine bond must pay, compared to the yield of 10-year bonds issued by the U.S. Treasury.

If the country risk is high, it will negatively affect the arrival of long-term investments. In addition, it generates problems in covering the financial needs of a country. Country risk thus relates to the eventuality that a sovereign state is unable to meet its obligations to a foreign agent, for reasons outside the normal risks arising from any credit relationship. Therefore, the JP Morgan country risk indicator (www.rava.com/perfil/riesgo%20pais, accessed on 18 December 2024) seems to be a reasonable proxy that captured investors’ risk perceptions about the political risk associated with the candidates during the 2023 Argentine presidential election.

In the equilibrium of the stock market, the stock price should be equal to the present value of the future cash flows or dividends expected by stockholders, while the discount rate reflects the required rate of return in consideration of the stock risk. Country risk generates a greater dispersion in cash flows and discount rates and, therefore, increases the volatility of returns. If political risk is not diversifiable, investors require a risk premium in compensation, and their returns are sensitive to such risk.

The above discussion seems to justify the application of a two-orthogonal-factor model by incorporating the innovations experienced in both market risk and country risk (as a proxy of political risk) and, once the impact of their respective sensitiveness on the stock returns of the Argentine industrial sectors is controlled, evaluating whether econometric estimations left space for the presence of abnormal returns either during the Argentine 2023 election process or after it.

3. Events’ Political Context and Research Design

3.1. Events’ Political Context

Argentina has fundamental electoral rules that were established in 1994 as a result of the constitutional reform that was carried out, and these rules were used in the elections of 1995, 1999, 2003, 2007, 2011, 2015, 2019, and the one studied in 2023.

Since 2009, candidates must first go through a primary known as PASO (Primary, Open, Simultaneous, and Mandatory). All candidates for the presidency of the country participate in these primaries, whether they belong to a political party or are independent. In the case of belonging to a party, all candidates who are competing internally for the leadership can participate in the primaries. Each candidate needs at least 1.5% of the votes to be able to run for president.

Once the candidates have been defined, the presidential elections are held. For a candidate to be elected president in the first round, he or she needs to obtain more than 45% of the votes, or 40%, provided that he or she is 10 percentage points ahead of the second-most-voted candidate.

Otherwise, a second round will be held in which the two candidates with the most votes will face each other in the next 30 days after the last election. A week before that date, a mandatory debate between the two candidates is held at the Faculty of Law of the University of Buenos Aires. In the second round, a simple majority is sufficient; in other words, the candidate who obtains the greatest number of valid votes in his or her favor will be elected. The presidential term in Argentina is four years, with the possibility of only one immediate re-election. However, a former president who has spent eight years in office can run again after a term away from power.

3.1.1. Event 1: Argentina 2023 Primary Elections (PASOs)

Primary elections or PASOs are mandatory elections that take place two and a half months before the general elections and are intended to select the candidates who will represent the political parties when they have two or more candidates for that nomination. In 2023, these elections took place on August 13 with the following candidates and results by political party:

• La Libertad Avanza (30.04%): Javier Milei.
• Juntos por el Cambio (28.27%): Patricia Bullrich, Horacio Rodríguez.
• Unión por la Patria (27.27%): Sergio Massa, Juan Grabois.
• Hacemos por Nuestro país (3.83%): Juan Schiaretti.
• Frente Izquierda (2.65%): Myriam Bregman.

The percentages of votes obtained by candidate were as follows:

With 30% of the votes, the candidate of the La Libertad Avanza party, Javier Milei, was surprisingly the winner of these primary elections, since he defeated the forces that governed Argentina for the last 20 years, the official Peronista–Kirchnerista coalition (Unión por la Patria), and Macrismo (Juntos por el Cambio), forces that in turn remained with the representation of the candidates Sergio Massa, obtaining second place overall with 21.4% of the votes, and Patricia Bullrich, obtaining third place overall with 17% of the votes (Statista.com 2023).

3.1.2. Event 2: Argentina 2023 General Elections (First Round)

After the primary elections, on Sunday, 22 October 2023, the general elections or first round were held. These elections had three parties (La Libertad Avanza, Juntos por el Cambio, and Unión por la Patria) positioned as the main ones to advance to the runoff, since under the applied surveys, none was a favorite by absolute majority. The favorite political party prior to this election was La Libertad Avanza, represented by the candidate Javier Milei, since it was the one that widely surpassed the second and third place in the PASO. However, these elections gave as the winner the Unión por la Patria party, represented by the candidate Sergio Massa, who won with 36% of the votes.

The percentages of votes obtained by each candidate were as follows:

With 36.6% of the votes, the representative of Unión por la Patria was proclaimed the winner of the 2023 Argentine general elections. However, these votes were not enough to achieve an absolute majority, corresponding to 45% of the votes, which would give way to the runoff election with the second most voted candidate. This election had Javier Milei (La Libertad Avanza) in second place with 29.98% and Patricia Bullrich (Juntos por el Cambio) in third place with 23.8% of the votes (Statista.com 2023).

3.1.3. Event 3: Argentina 2023 Runoff Elections (Second Round)

Since on 22 October 2023, none of the candidates managed to gather the sufficient percentage established by the Argentine constitution, the runoff election or second round was held. This election took place on 19 November 2023 and had the two candidates who obtained the most votes in the general election, who are Sergio Massa, 36.6%, and Javier Milei, 29.98%. Prior to the runoff elections, both candidates left their proposals open to Argentine society. Javier Milei’s proposals were as follows:

• Dollarization of the economy: Milei proposed dollarization as a measure to stabilize the Argentine economy and combat chronic inflation. This measure would eliminate the national currency (the Argentine peso) as a unit of account and completely replace it with the U.S. dollar.
• Exchange freedom: He advocated for the elimination of current exchange controls and allowing complete freedom in the purchase and sale of foreign currency. This would eliminate current restrictions and allow the market to determine the value of the dollar in relation to the Argentine peso.
• Reduction of public spending and elimination of taxes: He proposed significantly reducing public spending and eliminating taxes considered unnecessary or distorting, with the aim of encouraging investment and economic growth.
• Reform of the pension system: He proposed reforming the Argentine social security system to make it more sustainable and efficient, possibly incorporating elements of individual capitalization.
• Investment promotion: He sought to attract foreign and local investments by creating a more favorable regulatory environment and eliminating bureaucratic barriers.
• Deregulation and economic liberalization: He advocated for broad economic deregulation that allows companies to operate with fewer restrictions and encourages competition.
• Fight against corruption: He proposed implementing stricter measures to combat corruption at all levels of government and public administration.
• Reduction in the size of the State: Milei advocated for a significant reduction in the size of the state, arguing that this will encourage sustained economic growth and reduce bureaucracy.
• Tax reform: He proposed a simplification of the Argentine tax system, with the reduction of taxes for both individuals and companies, seeking to encourage investment and entrepreneurship.
• Labor flexibility: He advocated the flexibility of current labor laws, with the aim of making the labor market more dynamic and reducing structural unemployment.
• Economic deregulation: He proposed eliminating regulations that he considered unnecessary or that hinder economic activity, in order to promote competition and efficiency.
• Private education and health: He defended the expansion of private education and health as an alternative to the public system, arguing that this would improve the quality of services and offer more options to citizens.
• Institutional reforms: He proposed reforms in the judicial system and other state institutions to improve transparency and efficiency and to reduce corruption. He proposed implementing stricter measures to combat it at all levels of government and public administration.

These proposals reflect Milei’s focus on free market policies and his criticism of state interventionism, seeking to promote a more competitive Argentine economy that is less dependent on the state.

In addition, these proposals reflect a classic liberal stance that emphasizes economic freedom, the reduction of the size of the state, and the opening of markets as ways to improve Argentina’s economic situation, although they have generated controversy and debate in the country’s political spectrum.

On the other hand, Sergio Massa’s proposals were as follows:

• Economy and employment: Massa proposed economic policies that encourage national production, internal consumption, and job creation. He sought to promote investment in strategic sectors and support small and medium companies through tax and credit incentives.
• Education: He prioritized the improvement of educational quality and digital inclusion at all stages, from early education to technical and university training. He proposed programs to modernize school infrastructure and promote teacher training.
• Security: He proposed measures to improve citizen security, such as the creation of a judicial police force, the incorporation of technology in crime prevention, and the reform of the penitentiary system for social reintegration.
• Health: He focused on strengthening the public health system, improving the accessibility and quality of services, and promoting medical and scientific research in the country.
• Social development: He sought to implement social development policies that reduce poverty and inequality, such as the expansion of targeted social assistance programs and the creation of inclusive employment opportunities.
• Environment: He was committed to sustainable environmental policies and proposed promoting the use of renewable energy, the protection of natural resources, and the reduction of environmental pollution.
• Foreign policy: He advocated a foreign policy based on regional integration and the strengthening of diplomatic relations, with an emphasis on the defense of Argentine interests in the international arena.

These proposals reflect a view of Argentine economic, social, and environmental development, with the aim of improving the living conditions of Argentines and strengthening the country’s democratic institutions. In addition, the candidate had a strong focus on improving public services, such as education and health, and strengthening citizen security. One of Massa’s key proposals was to strengthen national production, implicitly supporting the industrial materials sector in Argentina.

3.2. Research Design

This article follows the classic event study steps (Henderson 1990) and includes important contemporary innovations to identify the best econometric method to measure stock market performance. The steps are as follows:

• Define the dates upon which the market would have received the news.

This article aimed to measure the effects of the announcements of the results of the presidential electoral process, which took place in 2023 in Argentina, on the stockholder returns of its most important industrial sectors. In this process, the following three key dates were important:

• Monday, 14 August 2023, corresponding to the day after the primary elections;
• Monday, 23 October 2023, corresponding to the day after the first general voting elections;
• Tuesday, 21 November 2023, corresponding to the day after the second voting elections, which ended with the winner of the Argentine presidency, the candidate of the political party “La Libertad Avanza”, Mr. Javier Milei.

The data include 187 daily excess returns (the stock return minus the daily risk-free rate associated with the Argentina Central Bank’s monetary policy) for each of the five industrial sectors under study, according to Figure 2.

2.

Characterize normal returns of the individual Argentine industrial sectors (financial, energy, public services, and other industries) in the absence of this news.

3.

Measure the difference between observed returns and “no news” returns for each Argentine industrial sector’s abnormal returns.

For steps 2 and 3, a multiple-index model (Henderson 1990) was used, based on two orthogonal risk factors (market risk and country risk) and on the deviations (abnormal returns captured by Jensen’s alphas) of an empirical version of Arbitrage Pricing Theory (Ross 1976), which is called throughout the article “a modified two factors model”. Then, four different econometric specifications were contrasted for it. The state-space method that allows time varying betas and residual conditional variance presents a better fit data compared to the remaining methods. Traditionally, previous researchers have added indexes for influences other than the market, as is the case in this study. For example, Langeteig (1978) and Thompson (1988), among others, added industry factors that were important for performance measurement.

4.

Aggregate the abnormal returns across Argentine industrial sectors and across time.

For each Argentine industrial sector, individual abnormal returns (the day after a specific electoral process: PASO primaries, first and second rounds of the presidential election) were examined, along with the aggregate abnormal returns, the day after the second round was over.

5.

Statistically test the aggregated returns to determine whether the abnormal returns are significant and, if so, for how long.

Individual and aggregate abnormal returns statistical tests were performed in order to evaluate their statistical significance. Also, statistical tests were performed within a post-election period to evaluate whether abnormal returns remained or disappear.

4. Data and Methodology

4.1. Data

Regarding the data used in the application of econometric estimation methods for the modified two-factor model, this article included a total of 21 companies that conform to the Argentine S&P Merval stock index, grouped into five industrial sectors:

The financial sector (six companies): Banco BBVA, Banco Macro, Bolsas y Mercados Argentinos, Grupo Financiero Galicia, Grupo Supervielle, and Banco de Valores;

The energy sector (four companies): Sociedad Comercial del Plata, Transportadora de Gas del Norte, Transportadora de Gas del Sur, and YPF;

The public services sector (four companies): Central Puerto, EDENOR, Pampa Energía, and TRANSENER;

The materials sector (three companies): Aluar Aluminio Argentino, Loma Negra Compañía, and Ternium Argentina;

The other industries sector (four companies): CRECUD, IRSA, Mirgor, and Telecom Argentina.

The S&P Merval is considered the flagship stock market index for Argentina. It seeks to measure the performance of the largest and most liquid stocks traded on “Bolsas y Mercados Argentinos (BYMA)”, which are classified as local stocks. Index components must meet minimum size and liquidity requirements. Their components (companies) are followed by thousands of investors who want to diversify their portfolios either locally or globally.

Table 2. Descriptive statistics.

Argentine Industrial Sector	Mean	Standard Deviation	Skewness	Kurtosis	JB	ADF
Financial	0.5842%	4.0857%	0.2597	4.4010	17.3959 ***	−11.35245 ***
Energy	0.4340%	4.5435%	2.3040	20.5428	2563.3420 ***	−10.55391 ***
Public services	0.4202%	3.9856%	0.7641	7.0082	143.3726 ***	−10.58979 ***
Materials	0.3141%	4.7650%	−0.3627	6.0443	76.3083 ***	−11.21250 ***
Other industries	0.4854%	4.4332%	1.7741	13.3966	940.2887 ***	−10.84175 ***
S&P Merval	0.4256%	3.8609%	0.5952	8.2208	223.4181 ***	−10.08096 ***

*** Significant at 1%.

presents the descriptive statistics of the excess daily stock returns of the five Argentine industrial sectors under study and the S&P Merval stock market index. Column JB reports the values associated with the Jarque–Bera normality test. Column ADF reports the values of the Dickey–Fuller augmented unit root test. The lags for this test were determined based on the Schwarz criterion. Source: own elaboration based on the Economatica^TM database.

Table 2 shows the basic descriptive statistics of the daily stock market excess return of each of the five Argentine industrial sectors under study from 16 May 2023 to 20 February 2024. The highest average daily excess return was achieved by the financial sector (0.5842%), followed by the other industries sector (0.4854%). The worst average daily excess returns were achieved by the materials sector (0.3141%) and the public services sector (0.4202%). The lowest risk, measured by the daily standard deviation of excess returns, was achieved by the public services sector (3.9856%), while the highest was achieved by the materials sector (4.7650%). One of the series of daily excess returns associated with the sectors reported in Table 2 shows negative bias, while four show positive bias; i.e., they present a distribution loaded with negative and positive excess returns. In addition, all exhibit leptokurtosis, with more pointed distributions and thicker tails than a normal distribution. The rejection of the null hypothesis of normality was confirmed in all the series under study when examining the Jarque–Bera test results. Conversely, the ADF test confirmed the rejection of the null hypothesis of the unit root, thus making all the excess returns series stationary. The leptokurtosis and the high presence of non-normality in the excess returns of the five series analyzed indicates that models with time-varying parameters accompanied by heteroscedasticity in the residual variance modeling can eventually show better indicators of information criteria or goodness of fit to the data. The results of this article will confirm this later.

4.2. Methodology

This section presents four different competing econometric specifications/estimation methods for a linear two-factor model ($r_{mt}$, ${CR}_t$) based on APT, which was modified to measure the presence or absence of abnormal stock returns in each of the five industrial sectors on the trading day following the results of the primary elections and first- and second-round elections during the 2023 Argentine presidential election process.

• Method 1: [OLS, which assumes constant systematic risk (beta) accompanied by a homoscedastic residual variance].

  $$r_{it}={c\left(1\right)}_i\ast d1+{c\left(2\right)}_i\ast d2+{c\left(3\right)}_i\ast d3+{c\left(4\right)}_i\ast d4+{c\left(5\right)}_i\ast {CR}_t+{{\beta }^{\prime }}_i{r}_{mt}+{\epsilon }_{it}$$

  (1)

  where

• $r_{it}={\displaystyle \frac{P_{it}}{P_{it-1}}}-1$ is the excess stock return of sector i on day t;
• $P_{it}$ is the stock price (adjusted by capital variation) of sector i on the day t;
• $P_{it-1}$ is the stock price (adjusted by capital variation) of sector i on the day t − 1;
• ${CR}_t$ is the daily percent variation of orthogonal country risk factor (orthogonal to $r_{mt}$ stock market factor), created from the JP Morgan country risk indicator for Argentina (used as a proxy of Argentine political risk innovations);
• ${{\beta }^{\prime }}_i$ is the beta coefficient or systematic risk of the industrial sector i within the complete estimation period of the parameters;
• $r_{mt}={\displaystyle \frac{I_{mt}}{I_{mt-1}}}-1$ is the excess stock return of the S&P Merval index on day t;
• $I_{mt}$ is the S&P Merval index (adjusted by capital variation) on the day t;
• $I_{mt-1}$ is the S&P Merval Index (adjusted by capital variation) on the day t − 1;
• ${c\left(1\right)}_i$ captures the daily average abnormal return of sector i within the complete estimation period, except the days of the three events, on which it takes a value of 0;
• $d1$ is a binary variable that takes a value of 1 every day of the estimation period, except the days of the three events, on which it takes a value of 0;
• ${c\left(2\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the primary elections;
• $d2$ is a binary variable that takes a value of 1 on the first stock market trading day after the results of the primary elections and 0 otherwise;
• ${c\left(3\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the first voting general elections;
• $d3$ is a binary variable that takes a value of 1 on the first stock market trading day after the results of the first voting general elections;
• ${c\left(4\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the second voting elections;
• $d4$ is a binary variable that takes a value of 1 on the first stock market trading day after the results of the second voting elections;
• ${c\left(5\right)}_i$ represents the sensitiveness coefficient of sector i excess return to Argentina’s country risk innovations, ${CR}_t$;
• ${\epsilon }_{it}$ represents the error term for the industrial sector i on day t.

• Method 2: [Ordinary Least Squares OLS, which assumes constant systematic risk (beta) accompanied by a heteroscedastic residual variance governed by a GARCH process (1,1)].

  $$r_{it}={c\left(6\right)}_i\ast d1+{c\left(7\right)}_i\ast d2+{c\left(8\right)}_i\ast d3+{c\left(9\right)}_i\ast d4+{c\left(10\right)}_i\ast {CR}_t+{{\beta }^{″}}_i{r}_{mt}+{\mu }_{i,t}$$

  (2)

  $${\sigma }_{it}^2=w_i+{\delta }_i{\mu }_{i,(t-1)}^2+{\gamma }_i{\sigma }_{i,(t-1)}^2$$

  (3)

  where

• $r_{it}$ was defined previously;
• ${{\beta }^{″}}_i$ is the beta coefficient or systematic risk of the industrial sector i in the complete estimation period of the parameters;
• ${CR}_t$ was defined previously;
• $r_{mt}$ was defined previously;
• ${c\left(6\right)}_i$ captures the daily average abnormal return of sector i within the complete estimation period, except the days of the three events, on which it takes a value of 0;
• $d1$ was defined previously;
• ${c\left(7\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the primary elections;
• $d2$ was defined previously;
• ${c\left(8\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the first voting general elections;
• $d3$ was defined previously;
• ${c\left(9\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the second voting elections;
• $d4$ was defined previously;
• ${c\left(10\right)}_i$ represents the sensitiveness coefficient of sector i excess return to Argentina’s country risk innovations, ${CR}_t$;
• ${\mu }_{it}$ represents the error term for the industrial sector i on day t;
• ${\sigma }_{it}^2$ is the error term conditional variance for industrial sector i on day t, which follows a GARCH(1,1) process;
• $w_i$ is the constant term for industrial sector i;
• ${\delta }_i$ is the weight assigned to the squared error with one lag for industrial sector i on day t, which measures the sensitivity of conditional volatility to market shocks;
• ${\mu }_{i,(t-1)}^2$ is the news about the prior-day volatility of industrial sector i captured on day t and measured as the squared residuals with one lag, derived from the mean equation (ARCH term);
• ${\gamma }_i$ is the weight assigned to the conditional variance of the error term for industrial sector i on day t − 1, measuring the persistence over time of conditional volatility after a market shock;
• ${\sigma }_{i,(t-1)}^2$ is the error term conditional variance for industrial sector i on day t − 1, forecasted from the past period [Generalized Autoregressive Conditional Heteroscedasticity (GARCH) term].

• Method 3: [State-space, which assumes systematic risk (time-varying beta) that changes according to a mean reversion process accompanied by a homoscedastic residual variance].

  $$r_{it}={c\left(11\right)}_i\ast d1+{c\left(12\right)}_i\ast d2+{c\left(13\right)}_i\ast d3+{c\left(14\right)}_i\ast d4+{c\left(15\right)}_i\ast {CR}_t+{{\beta }^{‴}}_{it}\ast r_{mt}+{\pi }_{i,t}$$

  (4)

  $${\beta ‴}_{it}={\beta ‴}_i+{\varphi }_i({{\beta ‴}_{i(t-1)}-{\beta ‴}_i)+\tau }_{it}$$

  (5)

  where

• $r_{it}$ was defined previously;
• ${\beta ‴}_{it}$ is the beta coefficient or systematic risk of industrial sector i on day t, which follows a mean reversion process given by Equation (5);
• ${CR}_t$ was defined previously;
• $r_{mt}$ was defined previously;
• ${c\left(11\right)}_i$ captures the daily average abnormal return of sector i within the complete estimation period, except the days of the three events, on which it takes a value of 0;
• $d1$ was defined previously;
• ${c\left(12\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the primary elections;
• $d2$ was defined previously;
• ${c\left(13\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the first voting general elections;
• $d3$ is a binary variable, as defined previously;
• ${c\left(14\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the second voting elections;
• $d4$ was defined previously;
• ${c\left(15\right)}_i$ was defined previously;
• ${\pi }_{it}$ represents the error term for the industrial sector i on day t;
• ${\beta ‴}_i$ is the average systematic risk (beta) over the entire period of parameter estimation;
• ${\varphi }_i$ measures the speed that the systematic risk (beta) takes to converge to its mean value;
• ${\beta ‴}_{i(t-1)}$ is the lagged value of the systematic risk (beta);
• ${\tau }_{it}$ captures the error term or residual of expression (5) for the time-varying systematic risk (beta) of industrial sector i on day t.

• Method 4: [State-space, which assumes systematic risk (time-varying beta) that changes according to a mean reversion process accompanied by a heteroscedastic residual variance governed by a GARCH(1,1) process].

  $$r_{it}={c\left(16\right)}_i\ast d1+{c\left(17\right)}_i\ast d2+{c\left(18\right)}_i\ast d3+{c\left(19\right)}_i\ast d4+{c\left(20\right)}_i\ast {CR}_t+{{\beta }^{⁗}}_{it}\ast r_{mt}+{\rho }_{i,t}$$

  (6)

  $${\beta ⁗}_{it}={\beta ⁗}_i+{\varphi \prime }_i({{\beta ⁗}_{i(t-1)}-{{\beta }^{⁗}}_i)+\tau \prime }_{it}$$

  (7)

  $${\sigma \prime }_{it}^2={w\prime }_i+{\delta \prime }_i{\rho }_{i,(t-1)}^2+{\gamma \prime }_i{\sigma \prime }_{i,(t-1)}^2$$

  (8)

  where

• $r_{it}$ was defined previously;
• ${\beta ⁗}_{it}$ is the beta coefficient or systematic risk of industrial sector i on day t, which follows a mean reversion process given by Equation (7);
• ${CR}_t$ was defined previously;
• $r_{mt}$ was defined previously;
• ${c\left(16\right)}_i$ captures the daily average abnormal return of sector i within the complete estimation period, except the days of the three events, on which it takes a value of 0;
• $d1$ was defined previously;
• ${c\left(17\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the primary elections;
• $d2$ is a binary variable, as defined previously;
• ${c\left(18\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the first voting general elections;
• $d3$ was defined previously;
• ${c\left(19\right)}_i$ represents the abnormal return of sector i produced by the announcement of the results of the second voting elections;
• $d4$ was defined previously;
• ${c\left(20\right)}_i$ was defined previously;
• ${\rho }_{it}$ represents the error term for the industrial sector i on day t;
• ${\beta ⁗}_i$ is the mean systematic risk (beta) over the entire period of parameter estimation;
• ${\varphi \prime }_i$ measures the speed that the systematic risk (beta) takes to converge to its mean value;
• ${\beta ⁗}_{i(t-1)}$ is the lagged value of the systematic risk (beta);
• ${\tau \prime }_{it}$ is the error term or residual of expression (5) for the time-varying systematic risk (beta) of industrial sector i on day t;
• ${\sigma \prime }_{it}^2$ is the error term conditional variance for industrial sector i on day t, which follows a GARCH(1,1) process;
• ${w\prime }_i$ is the constant term for industrial sector i;
• ${\delta \prime }_i$ is the weight assigned to the squared error with one lag for industrial sector i on day t, which measures the sensitivity of conditional volatility to market shocks;
• ${\rho }_{i,(t-1)}^2$ is the news about the prior day volatility of industrial sector i captured on day t and measured as the squared residuals with one lag, derived from the mean equation (ARCH term);
• ${\gamma \prime }_i$ is the weight assigned to the conditional variance of the error term for industrial sector i on day t − 1, which measures the persistence over time of conditional volatility after a market shock;
• ${\sigma \prime }_{i,(t-1)}^2$ is the error term conditional variance for industrial sector i on day t − 1, forecast from the past period (GARCH term).

On the other hand, it was also examined whether the eventual presence or absence of abnormal returns during the announcement of the results of the 2023 Argentina election were maintained in an ex-post period, after Javier Milei was elected president.

This was to evaluate the implications of the information (public announcements) and exogenous shocks (in this case, the political event) on investors’ behavior regarding the hypothesis of market efficiency, including risk aversion (general sentiment) and overreaction (De Bondt 1989).

For estimation purposes, an econometric analysis was carried out based on the same bases of the modified two-factor model presented before in the methodological section of this article, with the exception that, in this case, a single coefficient was considered in order to capture the stock market performance (daily abnormal average return) in each of the five industrial sectors under study. The period analyzed covered 10 December 2023 (the day that Javier Milei assumed the presidency of Argentina) to 28 June 2024, totaling 145 daily returns.

The results obtained, after employing the four estimation methods, confirm again that method four turns out to be the best in terms of the values reached by the information criteria.

5. Econometric Methods Discussion

Method 1 relies on traditional OLS estimation, which assumes fixed parameters and homoscedastic residual variance for the modified two-factor model estimated in this study. This method yields the best values for the information criteria, such as Akaike, Schwarz, and Hannan–Quinn, when the data sample for econometric estimation purposes is stable (generally, monthly data avoids spillover effects), but this is not achieved under unstable, volatile stock market conditions when high frequency data (daily or intraday data) are used (Ferson and Harvey 1991, 1993; Holmes and Faff 2004; Benson et al. 2007).

Method 2 also relies on traditional OLS estimation, which assumes fixed parameters but with the innovation of a heteroscedastic residual variance, governed by a GARCH(1,1) process for the modified two-factor model estimated in this study. This method yields the best values for the information criteria when the data sample for econometric estimation purposes present volatility-clustering phenomena and thus unstable, volatile stock market conditions when high frequency data are used (Bollerslev et al. 1992).

However, recently, the financial literature shows that the state-space specifications (Adrian and Franzoni 2009; Holmes and Faff 2008; Mamaysky et al. 2007, 2008; among others) that allow time-varying parameters and homoscedastic residual variance (method 3) or those that include heteroscedastic residual variance governed by a GARCH(1,1) process (method 4) provide more precise measurement of the stock market performance by controlling time-varying systematic risk along with volatility-clustering phenomena through econometric estimation processes. This is due to recent studies in the field that have demonstrated substantial improvements in systematic risk measurement by estimating time-varying betas from returns measured at a higher frequency than monthly (Hooper et al. 2007; Reeves and Wu 2013; Ortas et al. 2015; Santos Da Costa et al. 2019; Sandoval and Molina 2022; Sandoval 2023, 2024; among others). The unknown parameters in the state-space of methods 3 and 4 in this study are estimated by maximizing the following maximum-likelihood function, according to Harvey (1990):

$$\log L_i \left({\theta }_i\right)=-{\displaystyle \frac{T}{2}}\log \left(2\pi \right)-{\displaystyle \frac{1}{2}}{\sum }_{t=1}^T\log f_{i,t}\left({\theta }_i\right)-{\displaystyle \frac{1}{2}}{\sum }_{t=1}^T{\displaystyle \frac{v_{i,t}^2\left({\theta }_i\right)}{f_{i,t }\left({\theta }_i\right)}}$$

(9)

where ${\theta }_i$ is the hyperparameter vector. For method 3, this vector considers, for each industrial Argentine sector, the coefficients c(11), c(12), c(13), c(14) and ${\beta ‴}_i$, and ${\varphi }_i$, respectively. For method 4, it considers coefficients c(15), c(16), c(17), c(18) and ${\beta ⁗}_i$, ${\varphi \prime }_i$ and ${\sigma \prime }_{it}^2$, respectively. $v_{i,t}^2\left({\theta }_i\right)$ are the predictive residuals (${\tau }_{it}$ and ${\pi }_{it}$, respectively) of methods 3 and 4, respectively, and the variance of these $f_{i,t }\left({\theta }_i\right)$ is estimated using a recursive Kalman filter algorithm. The initial values for the hyperparameter vector are set according to (Wells 1996), which highlights the initial value of 0.5 for the coefficients ${\varphi }_i$ in each method, which capture how quickly the market beta coefficients (time-varying) return to their mean. In addition, an initial value of $e^{-1}$ is set for the variance of the residuals of the equations of methods 3 and 4. The results do not significantly change when performing a sensitivity analysis of these values.

As such, unlike static methods that assume homoscedasticity for the residual variance, it is a more realistic assumption that the residuals of Equations (2) and (6) follow a GARCH(1,1) conditionally heteroscedastic process. In addition, several recent articles have reported the time-varying nature of betas, such as those by (Sandoval 2023, 2024), among others. Therefore, the parameters of the GARCH(1,1) processes in Equation (8) for method (4) are estimated via an iterative process that consists of first estimating the hyperparameter vector assuming homoscedasticity and then generating the predictive residuals to model their conditional variance according to the GARCH(1,1) process in Equation (8). The state-space system is then estimated again, generating the final results of the estimates.

6. Main Findings

This section reports the values obtained from the different information criteria associated with each Argentine industrial sector once the estimation of the modified two-factor model was implemented using each of the four competing estimation methods (methods 1–4). In econometrics, an Information Criterion (Akaike, Schwarz, and Hannan–Quinn) is a method used to select the best model from a set of models by maximizing the likelihood of the data, while penalizing the number of parameters to prevent overfitting. The lower the information criterion value, the better the model. The results according to the industrial sector are detailed as follows.

According to the results reported in

Table 3. Information criteria: financial sector.

Estimation Method	Type	Residual Variance	Schwarz Criterion	Akaike Criterion	Hannan–Quinn Criterion
Method 1	OLS	HOM	−4.7843	−4.6806	−4.7423
Method 2	OLS	HET	−4.9783	−4.8228	−4.9153
Method 3	State-space	HOM	−4.9199	−4.7644	−4.8568
Method 4	State-space	HET	−5.0292	−4.8909	−4.9732

Source: own elaboration.

,

Table 4. Information criteria: energy sector.

Estimation Method	Type	Residual Variance	Schwarz Criterion	Akaike Criterion	Hannan–Quinn Criterion
Method 1	OLS	HOM	−5.3378	−5.2341	−5.2958
Method 2	OLS	HET	−5.4078	−5.2523	−5.3448
Method 3	State-space	HOM	−5.3340	−5.1785	−5.2710
Method 4	State-space	HET	−5.4269	−5.2886	−5.3709

Source: own elaboration.

,

Table 5. Information criteria: public services sector.

Estimation Method	Type	Residual Variance	Schwarz Criterion	Akaike Criterion	Hannan–Quinn Criterion
Method 1	OLS	HOM	−5.3199	−5.2162	−5.2779
Method 2	OLS	HET	−5.4447	−5.2892	−5.3817
Method 3	State-space	HOM	−5.2886	−5.1331	−5.2256
Method 4	State-space	HET	−5.4682	−5.3299	−5.4122

Source: own elaboration.

,

Table 6. Information criteria: materials sector.

Estimation Method	Type	Residual Variance	Schwarz Criterion	Akaike Criterion	Hannan–Quinn Criterion
Method 1	OLS	HOM	−4.3108	−4.2072	−4.2688
Method 2	OLS	HET	−4.4022	−4.2467	−4.3392
Method 3	State-space	HOM	−4.3156	−4.1601	−4.2526
Method 4	State-space	HET	−4.4385	−4.3003	−4.3825

Source: own elaboration.

and

Table 7. Information criteria: other industries sector.

Estimation Method	Type	Residual Variance	Schwarz Criterion	Akaike Criterion	Hannan–Quinn Criterion
Method 1	OLS	HOM	−4.5815	−4.4778	−4.5394
Method 2	OLS	HET	−4.7590	−4.6035	−4.6960
Method 3	State-space	HOM	−4.5494	−4.3939	−4.4864
Method 4	State-space	HET	−4.7513	−4.6131	−4.6963

Source: own elaboration.

, method 4 presented the most negative values for all the information criteria in the five sectors. The exception is the other industries sector, where method 4 is better, according to the Akaike and Hannan–Quinn criteria. This confirms that the state-space (SS) estimation method, which assumes a time-varying beta that changes according to a mean reversion process accompanied by a heteroscedastic (HET) residual variance governed by a GARCH(1,1) process, is a dominant econometric specification that better fits the data when estimating the modified linear two-factor model.

Table 8. Individual abnormal returns on the day after the announcement of the election stage results: method 4 results for financial, energy, public services, and material sectors.

Argentine Industrial Sector	Election Stage	Winning Candidate	Abnormal Returns	Parameter	Value	Sig
Financial	Primary	Javier Milei	Negative	C (17)	−4.8%	***
	First voting	Sergio Massa	Positive	C (18)	+10.7%	***
	Second voting	Javier Milei	Non-Significant	C (19)	−1.0%	-
Energy	Primary	Javier Milei	Positive	C (17)	+2.6%	***
	First voting	Sergio Massa	Negative	C (18)	−2.8%	***
	Second voting	Javier Milei	Positive	C (19)	+14.1%	***
Public services	Primary	Javier Milei	Positive	C (17)	+2.2%	***
	First voting	Sergio Massa	Non-Significant	C (18)	−0.9%	-
	Second voting	Javier Milei	Non-Significant	C (19)	−0.1%	-
Materials	Primary	Javier Milei	Positive	C (17)	+2.1%	***
	First voting	Sergio Massa	Negative	C (18)	−8.1%	***
	Second voting	Javier Milei	Negative	C (19)	−14.5%	***
Other industries	Primary	Javier Milei	Positive	C (7)	+0.6%	**
	First voting	Sergio Massa	Positive	C (8)	+5.7%	***
	Second voting	Javier Milei	Positive	C (9)	+10.8%	***

Source: own elaboration. *** Significant at 1%, ** significant at 5%.

shows the results of significant individual abnormal returns by the industrial sector and election stage in Argentina, and

Table 9. Individual abnormal returns on the day after the announcement of the election stage results and general candidates’ proposals through the electoral process.

Argentine Industrial Sector	Election Stage	Winning Candidate	Abnormal Return Sign	Massa’s General Proposals Announced Through Electoral Process	Milei’s General Proposals Announced Through Electoral Process
Financial	Primary	Milei	−	✓ National production ✓ Internal consumption ✓ Job creation ✓ Quality and digital education ✓ Educational inclusion in all levels ✓ Improve citizen security ✓ Strengthening public health system ✓ Poverty and inequality reduction ✓ Sustainable environmental policies ✓ Foreign policy based on regional integration	✓ Dollarization of the economy ✓ Exchange freedom ✓ Reduction of public spending and elimination of taxes ✓ Reform of the pension system ✓ Private investment promotion ✓ Deregulation and economic liberalization ✓ Fight against corruption ✓ Reduction in the size of the state ✓ Tax reform ✓ Labor flexibility ✓ Economic deregulation ✓ Private education and health ✓ Institutional reforms
	First voting	Massa	+
	Second voting	Milei	NS
Energy	Primary	Milei	+
	First voting	Massa	−
	Second voting	Milei	+
Public Services	Primary	Milei	+
	First voting	Massa	NS
	Second voting	Milei	−
Materials	Primary	Milei	+
	First voting	Massa	−
	Second voting	Milei	−
Other Industries	Primary	Milei	+
	First voting	Massa	+
	Second voting	Milei	+

NS: Non-statistically significant at any conventional statistical significance level. − and + sign values are reported in Table 8. All values are statistically significant at 1% level. Source: own elaboration.

presents the sign of individual abnormal returns on the day after the announcement of the election stage results, along with a summary of the general candidates’ proposals announced through the 2023 electoral process. In the case of the financial sector, the electoral victories of Javier Milei (who won the primary and second-stage voting elections) generated negative abnormal returns that were significant at 1% only in the primary elections. It seems investors were adjusting their expectations through the election process based on Milei’s economic plan to stabilize inflation and the Argentine peso devaluation (dollarization and free exchange rate determination). In the case of the energy sector, his victories generated positive abnormal returns in both the primary and second-stage voting elections, both significant at 1%. This result is also in line with Milei’s economic plan based on increased competition and efficiency (economic deregulation and liberalization) through the different Argentine industries, where the energy sector as a supplier is a key factor. In the case of the public services sector, they generated positive abnormal returns that were significant at 1% only in the primary elections. Investors were adjusting their expectations through the election process based on Milei´s economic plan oriented to increase competition and efficiency within this sector (private investments rather than public and public spending reduction). In the materials sector, this generated positive abnormal returns that were significant at 1% in the primary elections but negative and significant at 1% in the second-stage voting election. Investors were adjusting their expectations through the election process based on Milei’s economic plan to eliminate economic protection to this sector (economic deregulation and liberalization). To continue protecting this sector was part of Massa’s proposal (national production and internal consumption), but not in line with Milei’s economic plan. Finally, in the other industries sector, positive abnormal returns that were significant at 5% and 1% were observed in the primary, first-, and second-stage voting elections, respectively. This is also in line with Milei’s economic plan to increase competition, with more participants who can create value through their projects in this sector (economic deregulation and liberalization).

Table 10. Conjoint abnormal returns on the days after the announcement of the election stage results: method 4 results for financial, energy, public services, and material sectors.

Argentine Industrial Sector	Conjoint Null Hypothesis in Method 4	Wald Test p-Value	Abnormal Returns Sum	Hypothesized Effect	Main Finding Effect
Financial	c (17) + c (18) + c (19) = 0	0.2217	−4.9%	+	NS
Energy sector	c (17) + c (18) + c (19) = 0	0.0000	+14.0% ***	+	+
Public services sector	c (17) + c (18) + c (19) = 0	0.6541	+1.2%	−	NS
Materials sector	c (17) + c (18) + c (19) = 0	0.0000	−20.6% ***	−	−
Other industries	c (17) + c (18) + c (19) = 0	0.0000	+17.1% ***	+	+

Source: own elaboration. *** Significant at 1%, NS: Non-significant, +/−: Hypothesized positive/negative relationship between Milei’s general proposals announced through electoral process and conjoint abnormal returns on the days after the announcement of the election stage results.

shows the conjoint abnormal returns on the days after the announcement of the election stage results. Its fourth column presents the results with the sum of the individual abnormal returns associated with the business days immediately following the result announcements of the primary and first- and second-round elections in the presidential election process in Argentina in 2023. According to the results, statistically significant joint abnormal returns were obtained at 1% only in the energy, materials, and other industries sectors.

Abnormal returns reached +14%, −20.6%, and +17.1% in the energy, materials, and other industries sectors, respectively. Meanwhile, no statistically significant abnormal returns were observed in the financial and public services sectors.

These results validate our hypotheses for the energy, materials, and other industries sectors. Non-significant evidence was found to support our hypotheses for the financial and public services sectors once the electoral presidential election was over.

Table 11. Mean reversion sector market beta (systematic risk) and sensitiveness coefficient of excess return to Argentina´s country risk innovations: method 4 results for financial, energy, public services, and material sectors.

Argentine Industrial Sector	Mean reversion Sector Market Beta	Coefficient of Country Risk Sensitiveness
Financial	1.1221 ***	C(20) = −0.1069
Energy sector	0.9484 ***	C(20) = +0.1307 ***
Public services sector	0.9515 ***	C(20) = +0.1114 ***
Materials sector	0.8629 ***	C(20) = +0.1074
Other industries sector	0.8822 ***	C(20) = +0.0500

Source: own elaboration. *** Significant at 1%.

and Figure 3 show market betas (mean reversion systematic risk) and the sensitiveness coefficient of each sector excess returns to Argentina’s country risk. All market betas are significant at the 1% level. The financial sector shows the highest market beta risk, followed by the public services sector. The other industries and materials sectors, on the other hand, show the lowest market beta risks. Regarding the coefficient of country risk sensitiveness, the energy and public services sectors present the highest coefficients, both significant at the 1% level, while the remaining sectors are not significant at any conventional level of statistical significance.

Another objective of this article is to evaluate the five Argentine economic sectors’ stock market performance in a period after Milei’s election, in order to corroborate whether or not the abnormal returns obtained after the announcement of the results during the 2023 electoral process maintained sign and statistical significance during a period after the elections.

The above was performed to compare whether or not the stock market performance, once the 2023 electoral process was over, was aligned with the stock market performance shown within an ex-post period, with Javier Milei as president in Argentina (in office since 10 December 2023), and thus evaluate whether or not there was an under- or overreaction by investors in Argentina at the end of the electoral process in 2023.

The results obtained, after estimating the four estimation methods already indicated in the methodological section of this article, confirm again that method four turns out to be the best in terms of the values reached by the information criteria.

According to the results reported in

Table 12. Comparative abnormal returns on the days after the announcement of the election stage results versus ex-post abnormal average returns (10 December 2023 to 28 June 2024 period).

Argentine Industrial Sector	Wald Test p-Value	Abnormal Returns During Electoral Process	Abnormal Average Returns Post-Election Period	Z Test p-Value	Under/Overreaction
Financial	0.2217	−4.9%	0.31% **	0.0163	Underreaction
Energy sector	0.0000	+14.0% ***	0.10%	0.4620	Overreaction
Public services sector	0.6541	+1.2%	−0.18%	0.1058	-
Materials sector	0.0000	−20.6% ***	−0.18%	0.2138	Underreaction
Other industries sector	0.0000	+17.1% ***	−0.08%	0.2366	Overreaction

Source: Own elaboration. *** Significant at 1%, ** significant at 5%.

, the financial sector does not show statistically significant abnormal returns at the end of the second voting 2023 electoral process. However, a positive average daily abnormal return of 0.31% is observed, which is statistically significant at the 1% level, in the post-election period. The stock price fundamentals of banking institutions in Argentina reacted beyond those of equilibrium, in the post-election period, showing either better-than-expected net cash flows for their stockholders or a lower general appreciation of risk that led to a reduction in the cost of capital and thus to positive abnormal returns. The above is in line with the economic and financial stability plan to control inflation and depreciation of Argentine peso announced by Milei. Thus, their investors showed an underreaction at the end of the electoral process, one anomaly of the strong form of market efficiency hypothesis. In the energy and other industries sectors, statistically significant positive abnormal returns of 14.0% and 17.1%, respectively, are observed at the end of second voting 2023 electoral process. However, no statistically significant average daily abnormal return is observed in the post-election period. The fundamentals of the stock prices of companies in these sectors thus returned to their equilibrium levels, showing only normal returns. Again, this result is in line with the economic and financial stability plan announced by Milei, which focuses on a gradual expansion of the Argentine economy, making the energy sector a key sector that contributes to increase the number of Argentine companies within competitive market conditions. Thus, their investors showed an overreaction at the end of the electoral process, which resulted in excessive optimism regarding the expected net cash flows for their stockholders or a lower general appreciation of risk, which led to a reduction in the cost of capital and thus to positive abnormal returns that disappeared after the electoral process, one anomaly of the strong form of market efficiency hypothesis.

In the public services sector, the absence of statistically significant abnormal returns is observed at the end of second voting 2023 electoral process and also afterwards. Thus, their investors did not show an under- or overreaction regarding the fundaments that determine the companies’ stock prices equilibrium in the sector. This result is also in line with the economic and financial stability plan announced by Milei, which focuses on a gradual expansion of the Argentine economy based on more competitive market conditions. However, this sector did not present anomalies regarding the strong form of market efficiency hypothesis.

Finally, in the materials sector, there are statistically significant negative abnormal returns of −20.6% at the end of second voting 2023 electoral process. However, there is no statistically significant average daily abnormal return in the post-election period. The fundamentals of the stock prices of the companies in this sector in Argentina reacted in the period after the electoral process, aligning with those of equilibrium in determining the price of their stocks. This result is also in line with the economic and financial stability plan announced by Milei, which focuses on a gradual expansion of the Argentine economy, making this raw material supplier sector of the Argentine economy another key actor that contribute to increase competitive market conditions. Thus, their investors showed an underreaction at the end of the electoral process, which was subsequently corrected with a better expectation of net cash flows for their shareholders or with a lower general appreciation of risk that led to a reduction in the cost of capital and thus to higher returns, correcting the negative abnormal returns present at the end of the electoral process, another anomaly of the strong form of market efficiency hypothesis.

7. Results Discussion

There is a diverse body of literature that shows the impact of presidential election results on the stock market. This evidence focuses mainly on the impacts on the performance of stock indices, rather than on the industrial sectors that comprise them.

First, those studies that focus on stock indices tend to show that when presidential elections are won by candidates associated with political parties that strongly support economic and personal freedoms rather than state intervention, they cause more significant abnormal returns in the stock market.

In the case of the 2019 Indonesian presidential election, the election of President Jokowi-Ma’aruf, who made announcements of government policies aimed at encouraging business activity for the country, generated positive and statistically significant abnormal returns for the KOMPAS 100 stock index (Ratnaningsih and Widanaputra 2019).

Kumara and Fernando (2020), after studying 15 different events in the case of Sri Lanka between 2009 and 2018, conclude that the CSE stock index presented an overreaction to good political–economic news and an underreaction to bad news. Among the good news they classified were the end of the civil war in Sri Lanka, the removal of the state of emergency, the meetings of the commonwealth heads of government, the implementation of the information rights act, and the introduction of a new income tax structure. In bad news, they classified the strike at the Katunayake Free Trade Zone, ethnic tension in Dharga town, the report of the Committee of Public Enterprises on the issue of treasury government bonds, and the conflict in Theldeniya and the state of emergency declared for 10 days.

In the case of India, in 2019, the Bharatiya Janata Party, a party that sponsored changes that favor greater business and personal freedoms, returned to power after its first victory in 2014, after defeating the National Congress Party, which had been in power for 49 long years. This generated positive and significant abnormal returns in a sample of 31 companies that are part of the Bombay stock index. However, these were lower than those obtained after the first election in 2014 (Chavali et al. 2020).

Cázares and Luna (2022), after examining the electoral processes of 2006, 2012, and 2018 in Mexico, conclude that the increase in electoral uncertainty generates an increase in the IPC stock index returns’ conditional volatility, showing that the election of a left-wing candidate generates a reduction in the average returns of the index, while the election of a right-wing candidate increases it, promoting better stock market performance when the winner of the presidential election sponsors an economic plan that encourages private business activity with less state intervention.

Niederhoffer et al. (1970) examine the impact of the U.S. presidential election results on the Down Jones Index (DJI) from 1900 to 1968. They show that when the election was won by a Republican candidate, the day following the election, the DJI rose eight times and fell just one time. On the other hand, when the election was won by a Democratic candidate, the day following the election, the DJI rose four times and fell five times.

This historical evidence has not changed enough from the 1972 U.S presidential election to the most recent one in 2024. During this time period, there have been 15 presidential elections along with Gerald Ford, who served as the U.S president after Nixon’s resignation from office as a result of Watergate scandal. When the election was won or the office was gained by a Republican candidate, the day following the election, the DJI rose six times and fell four times. On the other hand, when the election was won or the office was gained by a Democratic candidate, the day following the election, the DJI rose two times and fell four times.

These results support a superior stock market performance when a Republican candidate is elected president. Riley and Luksetich (1980) also show evidence in the same direction.

It is important to note that Republicans advocate supply-side economics that primarily benefits businesses and investors by cutting taxes on businesses, allowing the economy to hire more workers, in turn increasing demand and growth and thus increasing revenues for a stronger economy that offsets the initial revenue losses over time. They also advocate the right to pursue prosperity without government interference through self-discipline, enterprise, saving, and investing.

Previous studies show evidence that supports the presence of a generally superior stock market performance (showing higher abnormal significant returns) when presidential elections are won by candidates who advocate supply-side economics without government interference. In addition, some stock markets seem to over- or underreact to the information content associated with good–bad news associated with political–economical events. However, there is little evidence about the effect of presidential election results on specific industrial sectors, especially in emerging stock markets.

In one comparative study for the U.S., Obradović and Tomić (2017) documented that Donald Trump’s victory in 2016 created positive abnormal returns on the military, financial, and energy industries. However, non-statistically significant abnormal returns were found for the electronic and health sectors.

Therefore, there are certain communalities between some U.S. and Argentine sectors, especially regarding the energy sector. Presidential candidates supported by political parties (Trump by the Republican party and Milei by La Libertad Avanza) that promote an economic similar agenda for this key supplier sector, with the final objective of increasing the number of companies and thus create more competitive and efficient market conditions, show in both cases abnormal returns at the moment that ends the electoral presidential process. For the resting sectors, the evidence is mixed.

Therefore, Milei’s economic proposals, mainly oriented toward greater liberalization of the economy and less state intervention in Argentina’s productive activities, were interpreted as good news in the energy sector companies and other industries (mainly focused on the services markets) but not in the materials sector, which has been characterized as being a sector protected by the state. The financial and public services sectors did not generate significant joint effects in terms of abnormal returns.

However, ex-post stock market performance evaluation shows aligned results only in the public services sector case. On the other hand, the financial and material sectors show an underreaction, and the energy and other industries sectors show an overreaction at the moment the 2023 Argentine the electoral process was finished. This shows that investors can either under- or overestimate the value of the fundamental variables that determine the companies’ stock prices due to economic and public plans announced by candidates during the election process. These kinds of reactions tend to normalize once the winning presidential candidate is under exercise, showing the different sectors only normal returns.

8. Conclusions

This study advances the understanding of dynamic econometric specifications used in political event studies by comparing four econometric estimation methods applied to a two-risk-factors model (market risk and country risk). This comparison has the purpose of identifying which best fit the relevant data in both the 2023 presidential election process in Argentina and during a post-election period, in order to corroborate whether or not the abnormal returns obtained once the 2023 electoral process was over maintained their sign and statistical significance during the next post-election period.

Our findings have several important implications. First, the results allowed us to conclude that the two-risk-factors model, under the state-space method 4 estimation, which assumes systematic risk (time-varying beta) that changes according to a mean reversion process accompanied by a heteroscedastic residual variance governed by a GARCH(1,1) process, best fits the data according to all the information criteria (the Akaike, Schwarz, and Hannan–Quinn criteria). This finding supports recent financial empirical evidence that more dynamic specifications for the modified two-risk-factors model allow better control of the time-varying nature of systematic risk and volatility-clustering phenomena in returns, reflected in more robust estimates from an econometric perspective.

Second, the abnormal returns accumulated in the three business days after the announcement of the results of the primary and first- and second-round elections show that the electoral victories of President Javier Milei caused an above-normal stock market appreciation in the case of the energy and other industries sectors and below-normal in the case of the materials sector, which shows an anomaly of the strong form of market efficiency hypothesis but not of the semi-strong one. The abnormal returns tend to disappear in the post-election period, showing, therefore, either an investor’s over- or underreaction during the electoral process.

Third, with regards to country risk, it was possible to identify a higher sensitiveness on stock market returns, to this risk in those Argentine sectors oriented to create more competitive industrial conditions through economic deregulation and by substituting public investments with private ones.

In terms of practical applications, the results obtained in this study suggest that those who anticipated the results of the different presidential election stages by liquidating positions in companies in the materials sector (a sector that was important in Massa’s proposal, but not for Milei’s economic plan) and investing more in companies in the energy and other industries sectors (which were important in the process to develop a more competitive and efficient economy) achieved a superior stock market performance when rebalancing their portfolios once the 2023 Argentine presidential election process was over. This can be interpreted as a good recommendation in terms of portfolio diversification and resource allocation for investors in Argentina at the moment that the presidential electoral process was over. After this, the recommendation is to invest less in those economic sectors that overreacted and more in those that underreacted, because the abnormal returns, either positive or negative, tend to disappear after the elections.

For policymakers, the post-election results aim to show that the economic plans of presidential candidates should be known in advance, with the greatest certainty and depth possible, while the electoral processes are still running, whatever the stages and time requirements that they demand. Thus, the informational content of the economic proposals of presidential candidates will be promptly incorporated into stock prices, avoiding the presence of abnormal returns that induce either an investor’s under- or overreaction, leading to undesirable wealth redistributions in the Argentine stock market once the electoral process is over. This situation disappeared during the post-election period in almost in all sectors, except in the case of the financial sector, which presented a positive, but low, average daily abnormal return, showing that its stock prices have not yet arrived at the equilibrium stock market conditions.

Looking ahead, the results and conclusions of this article should motivate future research developments in this field that use a dynamic event study methodology, with the objective of controlling the time-varying systematic risk and volatility-clustering phenomena found in this study and thus to eventually gain more communalities in comparative industrial sectors in different developing countries such as Argentina and, on the other hand, in developed countries such as the U.S. This would allow a broader generalization of results.