See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/344504548

The Effect of COVID-19 Pandemic on the Philippine Stock Exchange, Peso-

Dollar Rate and Retail Price of Diesel

Article  in  Journal of Asian Finance Economics and Business · October 2020

DOI: 10.13106/jafeb.2020.vol7.no10.543

CITATIONS READS

0 520

2 authors:

Aileen Camba Abraham Camba Jr.

Polytechnic University of the Philippines Polytechnic University of the Philippines
5 PUBLICATIONS   2 CITATIONS    5 PUBLICATIONS   2 CITATIONS   

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Aileen Camba on 22 October 2020.

The user has requested enhancement of the downloaded file.

 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553 543

Print ISSN: 2288-4637 / Online ISSN 2288-4645
doi:10.13106/jafeb.2020.vol7.no10.543

The Effect of COVID-19 Pandemic on the Philippine Stock Exchange,

Peso-Dollar Rate and Retail Price of Diesel

Aileen L. CAMBA1, Abraham C. CAMBA Jr.2

Received: July 18, 2020  Revised: August 23, 2020  Accepted: September 10, 2020

Abstract
This paper examines the effect of COVID-19 pandemic on the Philippine stock exchange, peso-dollar rate and retail price of diesel
using robust least squares regression and vector autoregression (VAR). The robust least squares regression using MM-estimation method
concluded that COVID-19 daily infection has negative and statistically significant effect on the Philippine stock exchange index, peso-
dollar exchange rate and retail pump price of diesel. This is consistent with the results of correlation diagnostics. As for the VAR model,
the lag values of the independent variable disclose significance in explaining the Philippine stock exchange index, peso-dollar exchange
rate and retail pump price of diesel. Moreover, in the short run, the impulse response function confirmed relative effect of COVID-19 daily
infections and the variance decomposition divulge that COVID-19 daily infections have accounted for only minor portion in explaining
fluctuations of the Philippine stock exchange index, peso-dollar exchange and retail pump price of diesel. In the long term, the influence
levels off. The Granger causality test suggests that COVID-19 daily infections cause changes in the Philippine stock exchange index and
peso-dollar exchange rate in the short run. However, COVID-19 infection has no causal link with retail pump price of diesel.

Keywords: Coronavirus, COVID-19 Pandemic, Philippine Stock Exchange, Robust Least Squares Regression, VAR Approach

JEL Classification Code: G11, G40, G41, I15, I18

1. Introduction valuations and increased volatility in the financial markets

affecting business investment, household consumption,
The novel coronavirus or COVID-19, which Chinese and international trade (Boon et al., 2020; Wren-Lewis,
authorities first reported to the World Health Organization 2020; Ramelli & Wagner, 2020). The Philippines, on
(WHO) on 31 December 2019, has spread globally. January 30, 2020, confirmed its first case of novel
The economic disruptions caused by the outbreak of coronavirus. Since then the virus has infected more than
COVID-19 is hurting economies, regardless of income 36,000 Filipinos (as of 30 June 2020), although more
level. The increased uncertainty is being reflected in lower than 1,200 individuals have recovered. The rapid rise in
domestic infections is disrupting the country’s economic
activities and value chains reducing demand nationwide.
First Author and Corresponding Author. Associate Professor
1 Moreover, the spread of COVID-19 to the provinces is
IV, Department of Economics, College of Social Sciences and expected to have a much larger impact on the Philippine
Development, Polytechnic University of the Philippines, Sta.
Mesa, Manila, Philippines [Postal Address: P. Tuazon Ave., Bgry. economy.
Kaunlaran, Quezon City, Metro Manila, 1111, Philippines] The uncertainty about the spread of the virus and the
Email: [email protected] level of aggregate demand could hurt the country’s stock
Assistant Professor II, Department of Economics, College of
2

Social Sciences and Development, Polytechnic University of the market, bring more volatility in the exchange rate, and
Philippines, Sta. Mesa, Manila, Philippines. could further depress oil prices. This study examines the
Email: [email protected] effect of the number of COVID-19 daily infections on
© Copyright: The Author(s) the Philippine stock exchange, peso-dollar rate and retail
This is an Open Access article distributed under the terms of the Creative Commons Attribution
Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0/) which permits price of diesel using robust least squares regression and
unrestricted non-commercial use, distribution, and reproduction in any medium, provided the
original work is properly cited. vector autoregression (VAR) approach.
544 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553

2.  Literature Review positive Average Abnormal Returns (AAR) during the
lockdown period. Since investors anticipated the lockdown,
2.1.  COVID-19 Impact on Stock Markets they reacted positively, however, in the pre-lockdown period
investors panicked with negative AAR. Khan et al. (2020)
Wren-Lewis (2020) claims that the COVID-19 pandemic investigates the impact of COVID-19 pandemic on the
would significantly affect the GDP of countries because of stock markets of 16 countries using pooled OLS regression,
reductions in production and consumer demands. Moreover, conventional t-test and Mann-Whitney test. Pooled OLS
if banks fail to meet the financial needs of firms due to estimation result shows that the growth rate of weekly new
falling demand, this will ultimately lead to the collapse of cases of COVID-19 negatively predicts the return in stock
the stock markets worldwide. Boon et al. (2020) describes markets. Using a t-test and Mann-Whitney test reveal that
three channels through which the COVID-19 pandemic investors in these countries do not react to the media news
may affect the global economy: 1) closure of factories, of COVID-19 at the early stage of the pandemic. However,
cutbacks in the service sector, disruption in the worldwide after human-to-human transmission has been confirmed,
supply chain will lead to an overall decline in the supply; stock market indices negatively reacted to the news in both
2) significant drop in travel and tourism, education and short- and long- event window.
other entertainment services will affect the demand side; 3)
increases in uncertainty will lead to a rise in the opportunity 2.2.  COVID-19 Impact on Exchange Rates
cost of investment. Ramelli and Wagner (2020) examines
the market reactions to the 2019 novel coronavirus disease The FRED Blog (2020) showed that COVID-19
(COVID-19) providing new insights into how real shocks pandemic has affected major economies’ exchange rates. In
and financial policies drive firm value. Thus, internationally- particular, Canada’s and the United Kingdom’s exchange
oriented firms, especially those more exposed to trade with rates spiked at the beginning of March 2020, but they seem
China, underperformed. As the virus spread to Europe and to have recovered by the end of March. The euro seems to be
the United States, corporate debt and cash holdings emerged unaffected by the global pandemic, compared with the U.S.
as important value drivers. Overall, the results illustrate how dollar. Despite the coronavirus origin, China’s exchange rate
anticipated real effects from the health crisis, a rare disaster, remained constant.
were amplified through financial channels. Corsetti and Marin (2020) concludes that the spread of
Gormsen and Koijen (2020) shows that news about fiscal COVID-19 has led to large foreign exchange movements.
stimulus boosts the stock market and long-term growth, but In particular, the pattern of foreign exchange dynamics is
did little to increase short-term growth expectations. Ozili fast-tracked and capital outflows from emerging markets,
and Arun (2020) found that restriction on internal movement week on week, are much larger compared to the previous
and higher fiscal policy spending had a positive impact on the 2007-2008 financial crisis. In ASEAN, Hayat and Mechelen
level of economic activities, although the increasing number (2020) sees currencies to weaken substantially in the
of confirmed coronavirus cases did not have a significant short term primarily due to COVID-19 pandemic. The
effect on the level of economic activities. Baker et al. (2020) uncertainties triggered a swift outflow of capital, causing
concludes that no previous infectious disease episode led to rapid depreciation of the exchange rates across the region
daily stock-market swings that even remotely resemble the (ASEAN Policy Brief, 2020). That is, key currencies in
response in the past month to COVID-19 developments. the region considerably depreciated, specifically the Thai
Yilmazkuday (2020) investigates the effect of the Baht, Indonesian Rupiah, and Singapore Dollar. The largest
coronavirus disease 2019 (COVID-19) cases in the U.S. depreciation was seen in the Indonesian Rupiah (IDR) at
on the S&P 500 Index using daily data covering the period 19.8 percent, from IDR13,662 per USD in end-January to
between December 31, 2019, and May 1, 2020. Using a IDR16,367 in end-March.
structural vector autoregression model, results suggest that Collins and Gagnon (2020) suggests that the most
having 1-percent increase in cumulative daily COVID-19 important policies to help economies from unwelcome
cases in the U.S. results in about 0.01 percent of a cumulative depreciations caused by COVID-19 pandemic includes
reduction in the S&P 500 Index after one day and about 0.03 providing access to central bank swap lines, beefing up
percent reduction after one month. Historical decomposition lending capacity at the international financial institutions
of the S&P 500 Index further suggests that the negative effect and keeping markets open for exports. Reserve-currency
of COVID-19 cases in the U.S. on the S&P 500 Index have countries may wish to consider coordinating direct foreign
been mostly observed during March 2020. Alam et al. (2020) exchange intervention with affected countries if these
investigates the impact COVID-19 to the stock market of unwelcome exchange rate tensions persist. Blanchard et
India from February 24 to April 17, 2020. The results al. (2015) provided considerable evidence that sterilized
confirm that the market reacted positively with significantly foreign exchange intervention does have significant effects
 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553 545

on exchange rates, especially in countries with restrictions largest oil exporter) to boost production to 12.3 mb/d, its
on cross-border capital mobility. Modest interventions full capacity. Saudi Arabia also announced unprecedented
between currencies of advanced economies with open discounts of almost 20 percent in key markets. The result
and sophisticated financial markets might have only was an immediate drop of more than 30 percent in prices
minor effects, evidence suggests that official purchases of and continuing declines since.
currencies of emerging-market and developing economies
often have more substantial effects. 3.  Research Methodology
2.3.  COVID-19 Impact on Oil Prices 3.1.  Robust Least Squares Regression
Pandemics such as COVID-19 can affect multiple The model for estimation takes into account the following
supply chains simultaneously and cover multiple regions functional forms:
globally. Hence, these effects last longer, they can be
more unpredictable, and they can result in simultaneous PSEI = f (COVID) (1)
disruptions in supply, demand, and logistics infrastructure.
For example, according to Ivanov (2020), 94 percent of PHUSD = f (COVID)(2)
the Fortune 1000 companies have experienced supply
chain disruptions due to COVID-19. Similarly, Araz et al. DIESEL = f (COVID)  (3)
(2020) underline that COVID-19 has broken many global
supply chains simultaneously. Although these effects are not where: PSEI is the Philippine stock exchange daily
unique to COVID-19, and they were also observed in earlier adjusted closing index, PHUSD is the Philippine peso-US
epidemic outbreaks works of Chou et al. (2004) and Calnan dollar daily closing exchange rate, DIESEL is the average
et al. (2018). retail pump price of diesel, and COVID the number of
Consistent with prior studies, Baldwin and Tomiura COVID-19 daily infections.
(2020) predicted that COVID-19 is both a supply and a For the convenience of parameter estimation, all
demand shock that will result in reductions in both imports variables have undergone logarithmic transformation:
and exports. These predictions are reflected in the paper LOG(PSEI), LOG(PHUSD), LOG(DIESEL and
of Yilmazkuday (2020) who investigated the COVID-19 LOG(COVID). Monthly data from January 31 to June 30,
cases from China and the rest of the world. Results based 2020, were sourced online from the websites of World
on a structural vector autoregression suggest that a weekly Health Organization (WHO), Yahoo Finance, Bangko
increase of 1,000 daily COVID-19 cases in China results Sentral ng Pilipinas (BSP) and the Philippine Department
in about 7 percent of a cumulative reduction in Baltic of Energy (DOE).
Exchange Dry Index after three months, whereas a weekly Before estimating the robust least squares regression,
increase of 1,000 daily COVID-19 cases in the rest of the it will be useful to established first the association of the
world results in about 0.4 percent of a cumulative reduction variables using correlation analysis. That is, coefficient
in crude oil prices. The historical decomposition results of correlation was utilized to examine the degree of
further suggest that the oil price plunge starting from association of the number of COVID-19 daily infections on
March 2020 is mostly explained by the corresponding the Philippine stock exchange index, Philippine peso-US
OPEC disagreement. Arezki and Nguyen (2020) identified dollar exchange rate and average retail pump price of diesel.
that negative supply and demand shocks have impacted The method of robust least squares regression is designed
oil prices. The demand component of the oil shock is to be robust, or less sensitive, to outliers. In this study we
linked to the sharp reduction in oil consumption stemming applied the method of MM-estimation (Yohai, 1987) to
from precautionary measures to stop the spread of the address any outliers in both the dependent (LOG(PSEI),
COVID-19. This includes lockdowns, which have brought LOG(PHUSD), LOG(DIESEL)) and independent
economies around the world to a standstill. In addition to (LOG(COVID)) variables. Thus, the regression models of
the shock from COVID-19, the breakdown in negotiations interest are:
between OPEC and its allies led to what will likely be a
persistent collapse in oil prices. On 5 March 2020, OPEC LOG(PSEI)t = α0 + α1 LOG (COVID)t + εt  (4)
proposed a 1.5 million barrel per day (mb/d) production LOG(PHUSD)t = α0 + α1 LOG (COVID)t + εt  (5)
cut for the second quarter of 2020, of which 1 mb/d would LOG(DIESEL)t = α0 + α1 LOG (COVID)t + εt  (6)
come from OPEC countries and 0.5 mb/d from non-OPEC,
most prominently Russia. The following day, Russia where: α0 and α1 are parameter estimates, ε the error term
rejected the proposal, prompting Saudi Arabia (the world’s and t represents time trend.
546 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553

3.2.  Vector Autoregression Model tested for the absence of Granger causality by estimating the
following VAR equations:
The vector autoregression (VAR) is used for forecasting
systems of interrelated time series and for analyzing yt = α0 + α1yt−1 + ⋅⋅⋅ + α1yt−1 + β1xt−1 + ⋅⋅⋅ + βlx−l + εt
the dynamic impact of random disturbances on the (8)
system of variables (D(LOG(COVID)) D(LOG(PSEI))
D(LOG(PHUSD)) D(LOG(DIESEL)). The reduced form xt = α0 + α1xt−1 + ⋅⋅⋅ + α1xt−1 + β1yt−1 + ⋅⋅⋅ + βly−l + ut  (9)
VAR approach sidesteps the need for structural modeling
by treating every endogenous variable in the system as for all possible pairs of (x, y) series in the group. The
a function of p-lagged values of all of the endogenous reported F-statistics are the Wald statistics for the joint
variables in the system (Christiano, 2012; Lütkepohl hypothesis: β1 = β2 = ⋅⋅⋅ = βl = 0 for each equation. The null
& Krätzig, 2004; Stock & Watson, 2001; Smith, 1993). The hypothesis is that x (COVID) does not Granger-cause y
stationary, k-dimensional, VAR(p) process can be written as (PSEI, PHUSD and DIESEL) in the first regression and that
y does not Granger-cause x in the second regression.
yt = A1yt−1 + ⋅⋅⋅ + Apyt−p + Cxt + εt  (7) Note that the variables are tested whether they are
stationary or not using Augmented Dickey–Fuller (ADF)
where yt is a kx1 vector of endogenous variables, xt (Dickey & Fuller, 1981) and Phillips-Perron (PP) (Phillips
is a dx1 vector of exogenous variables, A1, …, Ap are kxk & Perron, 1988).
matrices of lag coefficients to be estimated, C is kxd matrix
of exogenous variable coefficients to be estimated and εt is 4.  Results and Discussion
a kx1 white noise innovation process. This paper employs a
four-variable VAR system to investigate the causal relations 4.1.  Robust Least Squares Regression Results
and dynamic interaction of the Philippine stock exchange
The variables are tested first whether they exhibit
index (PSEI), the peso-dollar exchange rate (PHUSD) and
some degree of association. Paired correlations (r) were
the retail pump price of diesel (DIESEL) on the number of
calculated and further supported by the scatter diagrams
COVID-19 daily infections in the Philippines. and box plots in Figure 1. The corresponding correlation
We then developed impulse response function, variance patterns indicate some form of negative linear association
decomposition and Granger causality Wald test under VAR of the Philippine stock exchange index (PSEI), peso-dollar
framework with four endogenous variables. The impulse exchange rate (PHUSD) and retail pump price of diesel
response function (IRF) in the VAR framework was utilized (DIESEL) with respect to the number of COVID-19 daily
to trace the effect of a one-time shock to the innovations (i.e., infections (COVID). The correlation coefficients of -0.6215,
COVID) on current and future values of the endogenous -0.5014 and -0.6437 suggests strong negative association
variables. The variance decomposition give proportion of of PSEI, PHUSD and DIESEL with COVID, respectively.
the movements in the dependent variable that are due to their It is expected that increases in the number of COVID-19
own shocks versus shocks to other variables (Yuliadi, 2020; daily infections will lead to decreases in the Philippine stock
Brooks, 2019; Campbell, 1991). Variance decompositions exchange index, retail pump price of diesel and appreciation
are derived from a VAR with orthogonal residuals to facilitate of the Philippine peso against the United States dollar.
interpretation. Ordering the variables (Cholesky Ordering: Table 1 displays the results of the robust least squares
D(LOG(COVID)) D(LOG(PSEI)) D(LOG(PHUSD)) regression using MM-estimation method. Turning to the
D(LOG(DIESEL)) is important given the causal influence coefficient estimates, we see that the number of COVID-19
that they have on the dependent variable. The Granger (1969) daily infections has negative and statistically significant
(Pasquale, 2006) approach to the question of whether x effect on the Philippine stock exchange index, peso-dollar
(COVID) causes y (PSEI, PHUSD and DIESEL) is to see exchange rate and retail pump price of diesel. Since the
how much of the current y can be explained by past values calculated z-statistic values (in absolute term) of 9.0365,
of y and then to see whether adding lagged values of x can 4.6798 and 15.4535 are all significant at 1 percent level. Thus,
improve the explanation. y is said to be Granger-caused a one percent increase in the number of COVID-19 daily
by x if x helps in the prediction of y, or equivalently if the infections is associated with a 0.0227 percent and 0.0318
coefficients on the lagged x’s are statistically significant. In percent decrease in PSEI and DIESEL, respectively. Also, the
general, it is better to use more rather than fewer lags, since Philippine peso-US dollar exchange rate will appreciate by
the theory is couched in terms of the relevance of all past 0.0009 percent for every one percent increase in the number
information. In this study we picked a lag length of 8 that of COVID-19 daily infections. Noticed that the estimated
corresponds to reasonable beliefs about the longest time over coefficients have minimal values, however their negative
which one of the variables could help predict the other. We signs are in accord with the above correlation diagnostics.
 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553 547

Figure 1: Scatter Diagrams and Box Plots

Table 1: MM-estimation Method

Coefficient z-Statistic Probability

COVID-19 effect on the Philippine Stock Exchange Index
Constant 8.8760 463.4238 0.0000
LOG(COVID) -0.0227 -9.0365 0.0000

Adjusted R2 = 0.3892 Adjusted Rw2 = 0.5964

Rn2 = 81.6587    Prob. ( Rn2 statistic) = 0.0000
COVID-19 effect on the Peso-Dollar Exchange Rate
Constant 3.9296 2797.3860 0.0000
LOG(COVID) -0.0009 -4.6798 0.0000

Adjusted R2 = 0.1778 Adjusted Rw2 = 0.2448

2
Rn2 = 21.9006    Prob. ( Rn statistic) = 0.0000
COVID-19 effect on the Retail Pump Price of Diesel
Constant 3.6924 235.5324 0.0000
LOG(COVID) -0.0318 -15.4535 0.0000
2
Adjusted R2 = 0.4128 Adjusted R = 0.7862
w

2
Rn2 = 238.8095    Prob. ( Rn statistic) = 0.0000

S settings: tuning=1.547645, breakdown=0.5, trials=200, subsmpl=2, refine=2, compare=5. M settings: weight=Bisquare, tuning=4.684
Random number generator: rng=kn, seed=1314406110
Huber Type I Standard Errors & Covariance

The estimated equations also exhibit overall goodness-of- of the null hypothesis that all non-intercept coefficients are
fit and adjusted measures with R2 and Rw2 ranges from 0.1778 equal to zero. This implies that the transformed logarithmic
– 0.7862. It means that roughly 17.78 - 78.62 percent of the (log-log) models are significant. Thus, the number of
total variation in PSEI, PHUSD and DIESEL is accounted COVID-19 daily infections has significantly affected the
for or explained by the equations fitted on the given data. The Philippine stock exchange index, peso-dollar exchange rate
Rn2 statistic values of 81.6587, 21.9006 and 238.8095 with and retail pump price of diesel for the period January 31-
corresponding  p-value of 0.0000 indicate strong rejection June 30, 2020.
548 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553

Table 2: ADF and PP Unit Root Test Results

Levels First Difference

Order of
Variable Intercept, Linear Intercept, Linear
Intercept Intercept Integration
Trend Trend
Augmented Dickey-Fuller (ADF)
LOG(COVID) -1.3608 -0.6036 -7.4977*** -7.6053*** I(1)
(LOG(PSEI) -1.6250 -1.2730 -9.2341*** -9.3014*** I(1)
(LOG(PHUSD) -1.5057 -2.7455 -10.1884*** -10.1832*** I(1)
(LOG(DIESEL) -1.4128 -0.5066 -9.6972*** -10.1602*** I(1)
Phillips-Perron (PP)
LOG(COVID) -1.7863 -0.7306 -7.4977*** -7.6302*** I(1)
(LOG(PSEI) -1.8060 -1.5319 -9.4069*** -9.4373*** I(1)
(LOG(PHUSD) -1.5054 -2.8778 -10.1964*** -10.1836*** I(1)
(LOG(DIESEL) -1.4384 -0.5066 -9.6972*** -10.1596*** I(1)
Note: ADF test was performed using Schwarz information criterion and the automatic lag selection set at 14. PP test was performed with
Bartlett Kernel & Newey-West Bandwidth. Statistical sig.: ***(1%), **(5%) & *(10%).

Table 3: VAR Lag Order Selection Criteria

Lag LogL LR FPE AIC SC HQ

0 746.4992 NA 0.0000 -17.0690 -16.9556* -17.0233*
1 758.2438 22.1393 0.0000 -16.9711 -16.4043 -16.7429
2 766.7606 15.2715 0.0000 -16.7991 -15.7787 -16.3882
3 787.3627 35.0472 0.0000 -16.9049 -15.4310 -16.3114
4 808.4114 33.8715 0.0000 -17.0210 -15.0936 -16.2449
5 827.3468 28.7296 0.0000 -17.0884 -14.7076 -16.1297
6 844.6854 24.7126 0.0000 -17.1192 -14.2848 -15.9779
7 858.4536 18.3576 0.0000 -17.0679 -13.7800 -15.7440
8 882.6759 30.0690* 4.41e-13* -17.2569* -13.5155 -15.7504
9 895.7176 14.9905 0.0000 -17.1889 -12.9940 -15.4998
10 912.3089 17.5448 0.0000 -17.2025 -12.5541 -15.3308
* indicates lag order selected by the criterion AIC: Akaike information criterion
LR: Seq. modified LR test stat (each test at 5% level) SC: Schwarz information criterion
FPE: Final prediction error HQ: Hannan-Quinn information criterion

4.2  Vector Autoregression Model Results In Table 4, we can see the results of VAR model.
Checking what lag of D(LOG(COVID)) are significant
All the variables are stationary at first difference because in explaining D(LOG(PSEI)), D(LOG(PHUSD)) and
VAR model needs stationary data to run it as shown in Table D(LOG(DIESEL)), we refer to the system equation table.
2. The ADF and PP tests of the variables at levels are not In the system of equation, D(LOG(COVID)) lags 2, 3, 4,
stationary as they have unit root. However, the variables are 5, 6 and 8 are significant in explaining D(LOG(PSEI)) at 5
percent level of significance. When the dependent variable
not suffering from unit root when they are at first difference.
is D(LOG(PHUSD)), D(LOG(COVID)) lag 4 is significant
In Table 3, lag selection criterion, that is, LR, FPE and and when D(LOG(COVID)) is the dependent variable,
AIC suggests to take 8 lags in the VAR model to be optimum D(LOG(COVID)) lag 5 is significant in this system of
lags and employing inverse roots of AR characteristics equation. Moreover, the value of log likelihood function
polynomial indicates that the mod of reciprocal of each of 902.6691 for the VAR model is relatively large, and
characteristic root is inside the circle. That is, the lag order AIC value of -17.3184 is small, which indicates that the
of 8 is appropriate as reflected in Figure 2. explanatory ability of the model is very strong.
 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553 549

Figure 2: Inverse Roots

Table 4: Vector Autoregression Estimates

Dependent Variable
D(LOG(PSEI)) D(LOG(PHUSD)) D(LOG(DIESEL))
0.0101 -0.0017 0.0053
D(LOG(COVID(-1))) (0.0162) (0.0018) (0.0166)
[ 0.6236] [-0.9743] [ 0.3210]
-0.0363** -0.0013 0.0085
D(LOG(COVID(-2))) (0.0170) (0.0019) (0.0174)
[-2.1425] [-0.7089] [ 0.4862]
0.0583** 0.0006 0.0175
D(LOG(COVID(-3))) (0.0180) (0.0020) (0.0185)
[ 3.2479] [ 0.3162] [ 0.9458]
-0.0394** 0.0040** 0.0000
D(LOG(COVID(-4))) (0.0179) (0.0020) (0.0184)
[-2.2025] [ 2.0156] [-0.0007]
-0.0639** 0.0021 -0.0433**
D(LOG(COVID(-5))) (0.0176) (0.0019) (0.0181)
[-3.6382] [ 1.0702] [-2.3969]
0.0539** -0.0015 -0.0126
D(LOG(COVID(-6))) (0.0188) (0.0021) (0.0193)
[ 2.8679] [-0.7166] [-0.6522]
-0.0288 -0.0020 0.0136
D(LOG(COVID(-7))) (0.0201) (0.0022) (0.0206)
[-1.4340] [-0.9064] [ 0.6611]
0.0638** -0.0021 -0.0041
D(LOG(COVID(-8))) (0.0179) (0.0020) (0.0184)
[ 3.5618] [-1.0842] [-0.2202]
-0.0021 -0.0002 0.0032
C (0.0037) (0.0004) (0.0038)
[-0.5562] [-0.4710] [ 0.8290]
Log likelihood 902.6691
Akaike information criterion -17.3184
Standard errors in () & t-statistics in [ ] Significance level: **5%
550 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553

Figure 3: Response to Cholesky One S.D. (d.f. adjusted) Innovations

Table 5: Variance Decomposition

Variance Decomposition of Variance Decomposition of Variance Decomposition of

Day D(LOG(PSEI)): D(LOG(PHUSD)): D(LOG(DIESEL)):
S.E. D(LOG(COVID)) S.E. D(LOG(COVID)) S.E. D(LOG(COVID))
1 0.0305 8.5547 0.0031 0.0040 0.0301 0.0493
2 0.0312 9.7517 0.0032 1.0220 0.0304 2.1429
3 0.0317 10.5813 0.0032 2.1748 0.0310 2.0839
4 0.0318 10.6107 0.0032 2.2592 0.0310 2.1866
5 0.0318 10.6205 0.0032 2.2649 0.0310 2.1976
6 0.0318 10.6194 0.0032 2.2649 0.0310 2.1974
7 0.0318 10.6216 0.0032 2.2664 0.0310 2.1977
8 0.0318 10.6215 0.0032 2.2665 0.0310 2.1977
9 0.0318 10.6216 0.0032 2.2666 0.0310 2.1977
10 0.0318 10.6216 0.0032 2.2666 0.0310 2.1977
Cholesky Ordering: D(LOG(COVID)) D(LOG(PSEI)) D(LOG(PHUSD)) D(LOG(DIESEL))

Figure 3 shows the results of impulse response function infections will have impact on the Philippine peso-US dollar
under VAR. The blue line is the impulse response function exchange in the short run. Figure 3(c) demonstrated sharp
line while the red lines are simply the 95 percent coefficient decline from positive to negative, and then gradually tends
intervals. The impulse response function always lies within towards increment and becomes stable. It can be noticed
95 percent coefficient interval. We can observe the response that the number of COVID-19 daily infections has relatively
or the reaction of PSEI, PHUSD and DIESEL to the number affected retail pump price of diesel in the short term, and the
of COVID-19 daily infections. Figure 3(a) is the impulse effects lasts for 4 days.
response function of the Philippine stock exchange index Generally speaking, the number of COVID-19 daily
caused by COVID-19 daily infections shock. As shown in the infections has relatively affected the Philippine stock
figure, after a positive shock, the Philippine stock exchange exchange index, the Philippine peso-US dollar exchange
index tends to increase gradually and moves towards positive rate and retail pump price of diesel in the short term, but the
from negative and becomes steady at zero. Figure 3(b) is the effects are gentle in the long term.
impulse response function of the Philippine peso-US dollar In Table 5, we can see the results of variance decomposition
exchange rate to the number of COVID-19 daily infections. under VAR environment. Accordingly, in the short run (i.e.,
A one standard deviation shock or innovation of COVID-19 3rd day) shock to the number of COVID-19 daily infections
daily infections to the Philippine peso-US dollar exchange accounts for 10.5813 percent, 2.1748 percent and 2.0839
rate initially has noticeable impact, it increases from 1 to 2 percent variation of the fluctuations in the Philippine stock
days and then gradually decreases being in the negative zone exchange index, peso-dollar exchange rate and retail pump price
and then becomes stable. Thus, shocks to COVID-19 daily of diesel, respectively. In the long run (i.e., 10th day), variance
 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553 551

decomposition basically maintains stability at 10.6216 percent 5. Conclusions

for the Philippine stock exchange index, 2.2666 percent for
peso-dollar exchange rate and 2.1977 percent for retail pump This research strived to examine the effect of COVID-19
price of diesel. The magnitude of the contribution of the number pandemic on the Philippine stock exchange, peso-dollar
of COVID-19 daily infections does not change in the long run exchange rate and retail pump rice of diesel using robust
(i.e., 10th day). This means that the number of COVID-19 least squares regression and vector autoregression (VAR)
daily infections accounted for an insignificant portion of PSEI, approach. The paper established the correlation patterns
PHUSD and DIESEL fluctuations. suggesting strong negative association, that is, increasing
It can be seen from Figure 4 that, in the short term, number of COVID-19 daily infections will lead to decreases
the number COVID-19 daily infections has influenced the in the Philippine stock exchange index, retail pump price of
Philippine stock exchange index, the Philippine peso-US diesel and appreciation of the Philippine peso against the
dollar exchange and retail pump price of diesel. In the long US dollar. The robust least squares regression using MM-
term, the influence is leveling off. estimation method indicate that the number of COVID-19
Table 6 demonstrates the results of VAR Granger daily infections has significant but minimal effect on the
causality test using block exogeneity Wald test. We observed Philippine stock exchange index, peso-dollar exchange rate
that, when the Philippine stock exchange index is the
and retail pump price of diesel. Based on the VAR model,
dependent variable, the null hypothesis is rejected at 1 percent
the lag values of the independent variable are significant in
significance level. Also, when the dependent variable is the
explaining fluctuations in PSEI, PHUSD and DIESEL. A
peso-dollar exchange rate, the null hypothesis is rejected at
positive shock to COVID-19 daily infections has affected
10 percent level of significance. Therefore, it can be inferred
that in the case of the Philippines, the number of COVID-19 the Philippine stock exchange index, the Philippine peso-
daily infections causes changes in the stock exchange index US dollar exchange rate and retail pump price of diesel in
and peso-dollar exchange rate. That is, there is short-run the short term, but the effects becomes negligible in the
causality running from COVID-19 daily infections to the long term. Simply, the shock of the number of COVID-19
Philippine stock exchange index (COVID  PSEI) and daily infections differ in the short run and in the long run
peso-dollar exchange rate (COVID  PHUSD). However, to influence the fluctuations of the impulse receiver (or
COVID-19 daily infections have no causal link with retail dependent variables). We also discovered strong evidence
pump price of diesel in the short-run. That is, decreasing of causation running from COVID-19 daily infections to
prices of diesel for the period January 31-June 30, 2020, Philippine stock exchange index and from COVID-19 daily
could be attributed more to supply side shocks (Arezki & infections to peso-dollar exchange rate, not the other way
Nguyen, 2020). around.

Figure 4: Variance Decomposition Using Cholesky (d.f. adjusted) Factors ± 2 S.E.

Table 6: VAR Granger Causality (Block Exogeneity Wald Test)

Null hypothesis Chi-sq df Prob. Decision

D(LOG(COVID)) does not Granger cause D(LOG(PSEI)) 39.9698 8 0.0000 Reject
D(LOG(COVID)) does not Granger cause D(LOG(PHUSD)) 13.4144 8 0.0984 Reject
D(LOG(COVID)) does not Granger cause D(LOG(DIESEL)) 11.5831 8 0.1708 Accept
552 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553

References Collins, C. G., & Gagnon, J. E. (2020). Exchange rate policy in

the COVID-19 pandemic. Retrieved June 18, 2020, from
Alam, M., Alam, S., & Chavali, K. (2020). Stock Market Response https://www.piie.com/blogs/realtime-economic-issues-watch/
during COVID-19 Lockdown Period in India: An Event Study. exchange-rate-policy-covid-19-pandemic
Journal of Asian Finance, Economics and Business, 7(7), 131- Corsetti, G., & Marin, E. (2020). The dollar and international
137. https://doi.org/10.13106/jafeb.2020.vol7.no7.131 capital flows in the COVID-19 crisis. Retrieved May 12, 2020,
Araz, O. M., Choi, T. M., Olson, D. & Salman, F. S. (2020). from https://voxeu.org/article/covid-19-crisis-dollar-and-
Data analytics for operational risk management. Decision capital-flows
Sciences.  https://doi.org/10.1111/deci.12443 Dickey, D. A., & Fuller, W. A. (1981). Likelihood ratio statistics
Arezki, R., & Nguyen, H. (2020). Coping with a dual shock: for autoregressive time series with a unit root. Econometrica,
COVID-19 and oil prices. Retrieved June 18, 2020, from 49(1), 1057-1072.
https://voxeu.org/article/coping-dual-shock-covid-19-and-oil- Gormsen, N. J., & Koijen, R. (2020). Coronavirus: Impact on Stock
prices Prices and Growth Expectations. CEPR Policy Portal. Retrieved
ASEAN Policy Brief. (April 2020). Economic Impact of COVID-19 July 05, 2020, from https://voxeu.org/article/coronavirus-
Outbreak on ASEAN. Retrieved June 18, 2020, from https:// impact-stock-prices-andgrowth-expectations
asean.org/storage/2020/04/ASEAN-Policy-Brief-April-2020_
Granger, C. J. (1969). Investigating Causal Relationships
FINAL.pdf
by Econometrics Models and Cross Spectral Methods.
Baker, S. R., Bloom, N., Davis, S., Kost, K., Sammon, M., & Econometrica, 37, 425-435.
Viratyosin, T. (2020). The Unprecedented Stock-Market
Reaction to COVID-19. KelloggInsight. Chicago, IL: Kellogg Hayat, R., & Mechelen, R. (2020). How COVID-19 will
School of Management, Northwestern University. impact ASEAN: Deep recessions and a weak recovery.
Retrieved May 12, 2020, from https://economics.rabobank.
Baldwin, R., & Tomiura, E. (2020). Thinking ahead about the trade com/ publications/2020/may/ impact-covid-19-asean-
impact of COVID-19. London, UK: CEPR Press. Retrieved recessions-and-weak-recovery/#342881f1-72da-4673-b002-
June 18, 2020, from https://voxeu.org/content/economics-time- 70938b035513
covid-19
Khan, K., Zhao, H., Zhang, H., Yang, H., Shah, M. H., & Jahanger,
Blanchard, O., Adler, G., & Filho, I. (2015). Can Foreign Exchange
A. (2020). The Impact of COVID-19 Pandemic on Stock
Intervention Stem Exchange Rate Pressures from Global Capital
Markets: An Empirical Analysis of World Major Stock Indices.
Flow Shocks? IMF Working Paper No. 15/159. Retrieved July
Journal of Asian Finance, Economics and Business, 7(7), 463-
07, 2020, from https://www.imf.org/~/media/Websites/IMF/
474. https://doi.org/10.13106/jafeb.2020.vol7.no7.463
imported-full-text-pdf/external/ pubs/ft/wp/ 2015/_wp15159.
ashx Lütkepohl,  H., & Krätzig, M. (2004). Applied Time
Boon, L., Haugh, D., Pain, N., & Salins, V. (2020). Tackling Series Econometrics. Cambridge, UK: Cambridge University
the fallout from COVID-19. In Economics in the Time of Press. https://doi.org/10.1017/CBO9780511606885
COVID-19. London, UK: CEPR Press. Retrieved July 07, Foresti, P. (2006). Testing for Granger causality between stock
2020, from https://voxeu.org/content/economics-time-covid-19 prices and economic growth. MPRA Paper No. 2962. https://
Brooks, C. (2019). Introductory Econometrics for Finance. mpra.ub. uni-muenchen.de/2962/1/ MPRA_paper_2962.pdf
Cambridge, UK: Cambridge University Press. https://doi. Ozili, P., & Arun, T. (2020). Spillover of COVID-19: Impact on
org/10.1017/9781108524872 the Global Economy. Retrieved May 11, 2020, from https://doi.
Calnan, M., Gadsby, E. W., Konde, M. K., Diallo, A., & Rossman, org/10.2139/ssrn.3562570
J. S. (2018). The response to and impact of the Ebola epidemic: Phillips, P. C., & Perron, P. (1988). Testing for a unit root in time
towards an agenda for interdisciplinary research. International series regression. Biometrica, 75(1), 335-346. Retrieved June
Journal of Health Policy and Management, 7(5), 402. https:// 11, 2020, from http://www.jstor.org/stable/2336182
doi.org/10.15171/IJHPM.2017.104
Ramelli, S., & Wagner, A. F. (2020). Feverish Stock Price Reactions
Campbell, J. Y. (1991). A Variance Decomposition for Stock
to the Novel Coronavirus. Retrieved June 11, 2020, from
Returns. The Economic Journal, 101(405), 157-179. https://
https://doi. org/10.2139/ssrn.3550274
doi.org/10.2307/2233809
Christiano, L. J. (2012). Christopher A. Sims and Vector Smith, A. A., Jr. (1993). Estimating Nonlinear Time Series Models
Autoregressions. Scandinavian Journal of Economics, 114(4), Using Simulated Vector Autoregressions. Journal of Applied
1082-1104. https://doi.org/10.1111/j.1467-9442.2012.01737.x Econometrics, 8, 63–84. Retrieved June 11, 2020, from https://
www.jstor.org/stable/2285075
Chou, J., Kuo, N. F., & Peng, S. L. (2004). Potential impacts
of the SARS outbreak on Taiwanese economy. Asian Stock, J., & Watson, M. (2001). Vector Autoregressions. Journal of
Economic Papers, 3(1), 84-99. https://doi.org/10.1016/j. Economic Perspectives, 15, 101-115. https://doi.org/10.1257/
socscimed.2006.08.004 jep.15.4.101
 Aileen L. CAMBA, Abraham C. CAMBA Jr. / Journal of Asian Finance, Economics and Business Vol 7 No 10 (2020) 543–553 553

The FRED Blog. (2020). Coronavirus effects on exchange rates. Yilmazkuday, H. (2020). Coronavirus Disease 2019 and the
Retrieved June 11, 2020, from https://fredblog.stlouisfed. Global Economy. Retrieved June 11, 2020, from http://dx.doi.
org/2020/04/coronavirus-effects-on-exchange-rates/ org/10.2139/ssrn.3554381
Wren-Lewis, S. (2020). The economic effects of a pandemic. In: Yohai, V. J. (1987). High Breakdown-Point and High Efficiency
Economics in the Time of COVID-19. London, UK: CEPR Robust Estimates for Regression. Annals of Statistics, 15(20),
Press. Retrieved June 11, 2020, from https://voxeu.org/content/ 642-656. https://doi.org/10.1214/aos/1176350366
economics-time-covid-19 Yuliadi, I. (2020). An Analysis of Money Supply in Indonesia:
Yilmazkuday, H. (2020). COVID-19 Effects on the S&P 500 Index. Vector Autoregressive (VAR) Approach. Journal of Asian
Retrieved June 11, 2020, from http://dx.doi.org/10.2139/ Finance, Economics and Business, 7(7), 241-249. https://doi.
ssrn.3555433 org/10.13106/jafeb.2020.vol7.no7.241
View publication stats