Junctures in the time of COVID-19: Topic search and government's framing of COVID-19 response in the Philippines

Abstract

This article argues that, like many in Southeast Asia, the Philippine government's COVID-19 response was marked by policy experimentation and incremental adaptation, having been caught off-guard by the pandemic. Examining 16,281 government press releases related to COVID-19 issued by the Philippine News Agency between February 2020 and April 2021, we find that in its policy narratives the government panders initially to citizen demand, highlighting social amelioration as a pandemic strategy. However, as citizens’ economic anxiety further intensifies, the government's framing of the crisis response becomes pragmatic and turns towards promoting mass inoculation, ostensibly in a bid to convince citizens to choose health over short-term palliative economic measures. The findings nuance policymaking in an illiberal democracy, beyond the conventional populist description of seeking easy solutions or spectacularizing crisis response.

Just as prospects were getting rosy for the Philippine economy, bad news arrived. On January 22, 2020, the first COVID-19 infection in the Philippines was detected in two Chinese nationals visiting the country (Edrada et al., 2020). On February 2, 2020, the first confirmed death due to the virus was reported (Marquez, 2020). By March, the virus had taken the lives of 24 more additional patients and brought the number of cases to 380. By April that year, confirmed cases in the Philippines were running up to four digits.

The government's immediate response to the pandemic focused on minimizing its socioeconomic impact and restarting the economy. For instance, on March 23, 2021, Congress passed Republic Act 11469, the Bayanihan to Heal as One Act, granting the president, inter alia, temporary emergency powers to impose community lockdowns and providing cash assistance to displaced workers and low-income families. Essentially a “social amelioration initiative,” the Bayanihan Act was a direct response to the World Health Organization's (WHO) recommendations (Vallejo and Ong, 2020) to an economy thrown into its worst recession since the Second World War. While there were variations in subsequent implementing policies, within government the consensus seemed to have been towards balancing economic growth with public health. However, by the end of the first quarter of 2021, amidst a sudden surge in COVID-19 cases, the government's official policy response took a complete turn towards procuring ample vaccines and inoculating the public.

What causes policy response to turn from palliative interventions to pragmatic options? For resource-challenged governments conveying their policy narratives to the public in response to crises, what motivates the choice between economic viability and securing vaccines? Can citizens’ interest in an event or crisis spur policy learning amidst a lack of adequate opportunity to establish a response strategy?

This article argues that, like many in Southeast Asia, the Philippine government's COVID-19 response was marked by policy experimentation and incremental adaptation, having been caught off-guard by the pandemic. While it is well known in the public policy and political science literature that public opinion can shape public policy (Wlezien and Soroka, 2016); the bulk of the literature is usually in the context of interest groups, political parties, and elites, not ordinary citizens, and not with respect to experiences outside the USA (Burstein, 2003). As for the policy implications of the COVID-19 pandemic, on the other hand, the growing and emerging body of research is largely focused on epidemiological aspects such as the spread of the disease and compliance to health measures (Brodeur et al., 2021). There is also scant scholarly introspection on causal explanations pertinent to the governance component of pandemic policies beyond an organizing framework (Maor and Howlett, 2020).

Using the dynamics of the Philippine government's COVID-19 response as a case study, we argue that citizen concern on economic dislocations during the pandemic has a dynamic effect on how the government frames contentious policies whose redistributive implications can undermine government legitimacy. As anxious citizens seek information on the pandemic, the government panders to what it believes to be a response desired by the public and frames its COVID-19 policy around social assistance as a defining narrative. However, as mounting anxiety reveals the public's uncertainty with concurrent policy, the framing of the crisis response becomes rational and shifts towards the promotion of more viable options, such as mass inoculation. To support our conjecture, we analyzed 16,281 press releases on COVID-19 issued by the Philippine News Agency (PNA) between February 2020 and April 2021, and juxtaposed them with citizens’ COVID-19-related search interest based on Google Trends. Utilizing topic modeling and econometric techniques, we find that the government panders to citizen demand for economic safety nets in its policy narratives but only up to a point, and ultimately shifts to promoting mass inoculation as economic anxiety worsens.

The findings provide an evidence-based facet of policymaking in a precarious democracy criticized for lackluster pandemic intervention. The international media, for instance, initially branded the Philippine COVID-19 response a “tragedy of errors” (Beltran, 2020). While there is arguably a basis for this view, our findings suggest nevertheless that there was an attempt by the government to be responsive, even if such responsiveness meant pandering to public sensitivities with respect to weighing policy alternatives. There was also policy learning, as shown by how the government pragmatically reoriented its policy narrative from economic support to mass inoculation, after heightened interest in social transfers made it realize that population immunity is more cost effective in the long run.

The article proceeds as follows. We begin by expounding on the relationship between crises and policy change as the focus of theoretical and empirical introspection in the literature. Afterwards, we elaborate on our theory explaining why the government's COVID-19 policies were shaped by policy adaptation and experimentation. We also explain why we expect informational citizenship to induce the government to turn from a reactive policy that is narrowly focused on providing economic assistance to one that targets population immunity. This is followed by a discussion of the data, measures, and estimation approach we employed to test our conjecture. We then discuss the results and how they nuance current understanding of the government response to the COVID-19 crisis. The article concludes with the broader implications of the findings on pandemic policymaking.

Policymaking and policy shifts

There is a policy change or shift when one that is existing is replaced with something new and innovative, or undertakes incremental refinements (Bennett and Howlett, 1992). Political dynamics, problems, and proposal may be construed as streams capable of suddenly changing government policies (Cairney and Jones, 2016). During crucial periods, these streams may create an opportunity or condition in which agenda change is possible. Referred to as policy windows, these critical junctures open an array of choices to policy actors, allowing decision making to become more significant (Capoccia and Kelemen, 2007; Donnelly and Hogan, 2012). Punctuated equilibrium occurs when abrupt and radical policy changes arise after a long period of stability (Jones and Baumgartner, 2012). Bennett and Howlett (1992), for their part, argue that states learn from experiences and modify decisions based on how well previous measures performed in the past. This policy learning serves as a platform for future decisions and these are circulated as new information to achieve political goals and create policy-related beliefs over time (Moyson et al., 2017).

There are competing views on policy learning but there is almost scholarly consensus regarding the role of social forces in policy change. An indication of learning happens when there is a shift in policy, even though the preferred policy is not necessarily the most efficient (Hall, 1993). Sometimes the government learns simultaneously while responding to circumstances or crises through key societal actors who create conditions that state officials must address (Bennett and Howlett, 1992). Policy learning, according to Greener (2002), takes place at various stages—those that take place at the level of policy instruments, and those that involve a shift where policymakers reject their own ideas and adopt another. Whereas policymakers are often responsible for the former, the broader social and political forces are the driving factors behind paradigmatic shifts (Berman, 2013).

A robust strand of literature also construes policy change as a product of the confluence of three elements—institutions, interests, and ideas (Béland, 2009; Shearer et al., 2016). Institutions refer to the formal and informal rules, norms, beliefs, and precedents that shape the policy response to a crisis (Mahoney, 2000). Interests, on the other hand, reflect the policy choices, material motivations, and agenda, not just of policy actors (elected officials, civil servants), but even of those operating outside the government (societal groups, researchers, policy entrepreneurs) and their desire to influence the policy process to attain their own ends (Kern, 2011; Prechel and Boies, 1998). Meanwhile, ideas refer to knowledge or belief about what is (scientific or factual knowledge), what ought to be (values), or a combination of these (Pomey et al., 2010). Ideas can become decisive causal factors in policy change if there are considerable institutional impediments that weaken the capacity of political actors to promote the adoption of a concrete policy alternative (Béland, 2009).

However, institutions, ideas, and interests are not mutually exclusive. Policy change can generally be caused by a shift in one, or their combination, and alter the structure of a policy network. For ideas to be realized, for instance, there have to be advocates. Interests, on the other hand, need to be able to modify behavior to attain the survival of ideas by contending or cooperating with other interests. In the context of the policy process, ideas may reflect ideological considerations or citizens’ values. Ideology and values, in turn, are gauged empirically by looking at how public opinion plays a role in the diffusion or modification of policies. Wlezien (1995), for instance, conceives a model of public responsiveness in which citizens behave like a thermostat that sends a signal to the government when policy deviates from their preference so that this policy can change accordingly. Supposedly, critical public opinion ceases once policy is in congruence with citizen demand.

COVID-19 response as policy adaptation and experimentation

Existing works on the politics of the COVID-19 response in the Philippines have ubiquitously highlighted policy failure in the backdrop of a populist regime (Aguilar, 2020; Hapal, 2021; Teehankee, 2021). However, such preoccupation with populism as a one-size-fits-all account not only hampers a more nuanced examination of policy discourse (De Cleen and Glynos, 2021) but ignores the relevance of citizens as stakeholders and policy participants (Kweit and Kweit, 2004). This article digresses from the conventional populist formula and contributes through an empirical inquiry of the relationship between citizen demand and policy response.

Like many in Southeast Asia during the early phase of the pandemic, the Philippines’ response to COVID-19 was fidgety, imperfect, and disproportionate. Dewi et al. (2020) note that public officials with very limited information dealt with the pandemic through a plethora of policy strategies that were not necessarily the most effective. These disproportionate policy reactions are a form of policy overreaction driven by the challenges that elected officials faced while coordinating extant policy strategies and collecting public health information. In search of the most effective response, governments repeatedly undertake a trial-and-error process of agenda setting, formulation, implementation, and evaluation, while exploring multiple policy possibilities. At crucial junctures, policy problems, their solutions, and the political environment converge and pave the way for an intervention that is acceptable to the public. But because resources vary by country, governments also take different approaches in their struggle with the pandemic. Inevitably, they may also be constrained by balancing the need to secure citizens’ economic well-being with managing shocks to the healthcare system.

Studies such as those by Purnomo et al. (2022: 2) characterize the Philippines’ COVID-19 response as underreactive, or one in which there is “a systematic stagnant or inadequate response by government officials to high risks or no response at all.” Interestingly, countries which underreact to the pandemic also exhibit rapid COVID-19 transmission. Policy reaction, however, is not just a function of economic or institutional readiness. Dewi et al.'s (2020) findings imply, for instance, that overall collective opinion and well-being is related to the manner by which policymakers respond to the pandemic.

Our theoretical premise construes citizens’ pocketbook assessment of economic uncertainty under COVID-19—operationalized by their search interest on the pandemic—as catalyzing a critical juncture, and applies it in the context of the COVID-19 response in the Philippines (Figure 1). In our view, this juncture is capable of triggering a learning or realization process on the part of policymakers that allows incremental or small changes in policy through problem-solving approaches (Dunlop and Radaelli, 2013; Flink, 2017) and is reflected in policy narratives. We follow Moyson et al. (2017) and construe this learning process as having both cognitive and social components in which information and experience are used to inform, substantiate, or legitimize policy beliefs and objectives (Bennett and Howlett, 1992; Dunlop and Radaelli, 2013).

Figure 1.

Public's search interest and government's shifting policy narrative on COVID-19.

In this framework, stakeholder and citizen engagement feedback are valuable sources of policy learning, especially in the changing governance setting of governments grappling with COVID-19 as a critical juncture. We consider critical juncture as “a situation of extreme challenge and uncertainty” where institutional and social policies and practices have the propensity to result in fundamental change (Twigg, 2020). We contribute by providing empirical support to the notion of crisis as critical juncture and as a causal determinant of policy shifts, using the Philippine government's COVID-19 policy response as a case study.

As is perhaps the case elsewhere, policy response to COVID-19 in the Philippines was relegated to a specialized technical group that operated autonomously and made policy decisions that were often insulated from political factors. Accordingly, policy actions concerning COVID-19 routinely proceeded from a single source and had been, to borrow from the punctuated equilibrium literature, “in a relative state of equilibrium” (Amri and Drummond, 2020). The Inter-Agency Task Force for the Management of Emerging Infectious Diseases (IATF-EID) is the body primarily responsible for all matters concerning emerging infectious diseases, such as swine flu and the novel coronavirus. The IATF-EID was convened in January 2020 but it was neither ad-hoc nor new, having been created through Executive Order 168 issued by former president Benigno Simeon Aquino on May 26, 2014.

Since COVID-19 was declared a pandemic, the government's response had been highly reactive and focused initially on minimizing the novel coronavirus’ impact on the economy. For instance, three of the four pillars of the Duterte administration's socioeconomic strategy against COVID-19 (see e.g. Department of Finance, n.d.) were all towards providing short-term financial or monetary support to facilitate economic recovery. Building public health resilience as a long-run priority was never on the agenda and state officials reportedly even downplayed the threat of the virus (Dancel, 2020). The IATF provided operational command to the government's pandemic response but a coordination problem mired these efforts. Some local governments had to be more resourceful and took the initiative to secure vaccine doses instead of waiting for the national government to finalize agreements with international vaccine suppliers (Ranada, 2021). The IATF, as a government-led agency, was also criticized for its lack of appropriate pandemic experts (e.g. epidemiologists), its reliance on former military officers for leadership, and for discounting inputs from important stakeholders such as businesses (Inquirer, 2021).

By March 2021, the Philippines had had the world's longest lockdown—a notoriety largely derided in the international press as testament to the government's botched pandemic response (Madarang, 2021). Perhaps the government thought that by limiting people's mobility it would be able to curb the spread of the virus and make a quick rebound. But people had a different appreciation of the pandemic and held different perspectives as to which policies have the most impact. Community lockdowns dislocated people from their livelihoods, creating the very conditions the government wanted to avert. As the lockdown prolonged, citizens became weary of the existing COVID-19 response and started to demand alternatives that would ease economic dislocations during and beyond the pandemic. Community quarantines and restricted mobility soon shifted public discourse on government policy from economic safety nets to the benefits of mass inoculation. On February 28, 2021, the Philippines received its first COVID-19 vaccines—the last among ASEAN countries. The government officially began its vaccine rollout on March 1, 2021, starting with medical frontliners.

Data, variables, and analytical approach

Political texts are known to be vectors of policy positions (Laver and Garry, 2000). They also serve as a rich source of policy narratives which have played a key role in effective COVID-19 responses by creating opportunities for policy learning (Mintrom and O’Connor, 2020).

To gauge the Philippine government's policy position on COVID-19, we compiled a dataset consisting of a corpus of 16,281 press releases on the novel coronavirus posted online by the PNA between January 30, 2020 and April 15, 2021. Since the corpus consists of the government's own press releases, it is not expected to be disinterested or neutral. As many of the press releases are likely propaganda, their selective presentation of government programs may be more towards reinforcing personality politics in the Philippines than anything else. However, as is typical in studies of this type, our interest is precisely to read the government's policy position from a corpus consisting of its own statements. This is consistent with studies that analyze political texts from unilateral sources such as party manifestos (Eder et al., 2017), party elite interviews (Ecker et al., 2022), and public pronouncements or speeches by key political actors such as chief executives (Kaufman, 2020; Panao and Pernia, 2022).

The PNA (https://www.pna.gov.ph) is a web-based newswire service of the Philippine government, supervised by the News and Information Bureau (NIB) of the Presidential Communications Operations Office (PCOO), and is responsible for disseminating relevant newspaper articles to local and international news agencies. The data coverage begins with January 30 as this was the date of the first reported case of the novel coronavirus in the Philippines (Edrada et al., 2020). However, data-wise, the earliest PNA article on COVID-19 appeared only on February 15, 2021. We then derived topic models using Latent Dirichlet Allocation (LDA), a semi-supervised probabilistic approach of finding latent or hidden topics in a collection of documents by going through the distribution of co-occurring words or vocabulary (Blei and Jordan, 2004; Blei and Lafferty, 2007), and implemented them through the R-based application Quanteda (Benoit et al., 2018).

This procedure produced 10 topics. However, only seven of these exhibited thematic coherence. The topics contained words associated with the following: the economy, vaccination, testing, security, financial aid, hospitalization, and tourism. The procedure allows for the computation of a topic score (a value between 0 and 1) which may be construed broadly as the percentage by which a topic or theme characterizes a particular document. From this, we computed a mean topic score for each news date and constructed a day-to-day series, since COVID-19 updates occur on a daily basis. We operationalize the government's policy shift as the difference between the average daily scores for the economy and vaccination, respectively.

However, politicians and bureaucrats do not have a monopoly of policy formulation. We gauge citizen interest by examining the magnitude and pattern by which Filipinos have queried the COVID-19 pandemic on the internet. The big data platform Google Trends was utilized to construct a series of daily COVID-19-related searches for the Philippines categorized into the following search topics: economy (“economy,” “recession,” “unemployment,” “Philippine economy, “beneficiary,” “social amelioration program”), vaccination (“vaccine,” “COVID-19 vaccine,” “Astrazeneca,” “Pfizer,” “Moderna,” “Sinovac”), symptoms (“COVID-19 sign,” “COVID-19 symptom,” “asymptomatic”), and lockdown (“lockdown,” “community quarantine”). Although the categories are subjective, the search keywords are not arbitrary. Words such as unemployment, recession, and social amelioration, for instance, are all related search terms that trended with economy as a search topic. COVID-19 vaccine brands, particularly AstraZeneca, Pfizer, Moderna, and Sinovac, also figured prominently as popular search topics when people keyed the term “COVID-19 vaccine.” These are the four brands officially administered to the Filipino public as part of the government's inoculation effort at the time of data collection.

In recent years, online political discourse has come to be accepted as an approximation or surrogate of public opinion comparable to social surveys (Caldarelli et al., 2014; Kalampokis et al., 2013; Kwak and Cho, 2018; O’Leary, 2015). Aside from Twitter, Google Trends also figures in the literature as a big data alternative for gauging public opinion and for analyzing the way by which consumers seek information (Jun et al., 2018). Kwak and Cho (2018) believe that Google Trends reflects more genuine thoughts because it is not shared by other social media users and is less susceptible to the bandwagon effect and social desirability bias. In practice, Google Trends is also relatively easier to use.

In the politics and policy literature, Google Trends has figured as a measure of issue salience (Dube and Kaplan, 2012; Graefe et al., 2014; Reilly et al., 2012) as well as public attention to issues (Ciuk and Yost, 2016; Weeks and Southwell, 2010). It has been used to study energy policy (Oltra, 2011), public interest on space exploration (Whitman Cobb, 2015), public interest on biodiversity (Troumbis, 2017), the impact of local health restriction on abortion (Reis and Brownstein, 2010), and the impact of global public health on health-seeking behavior (Havelka et al., 2020), and even to predict presidential elections (Prado-Román et al., 2021) and construct an index of racial prejudice (Stephens-Davidowitz, 2014).

We clarify that resort to search trends is our attempt at approximating public opinion. By search trend, we refer specifically to search interest—that is, what Filipinos search for in real time as events unfold—and not citizens’ news consumption. A survey conducted in 2021 by Pulse Asia indicates that about six in 10 Filipinos use the internet and also about six in 10 log in more than once a day (Gonzales, 2021a). Admittedly, search interest is a crude measure, considering that the same survey also indicates that television remains the top news source for Filipinos, especially on politics and government. Only 48 percent of Filipinos rely on the internet for news, of which 44 percent say they get their news from Facebook (Gonzales, 2021b). This is a limitation we acknowledge in this article.

We construe the aforementioned COVID-19 search trends as encompassing the Filipino public's concern or interest in COVID-19-related issues. We hypothesize that in the framing of its COVID-19 policy, the government's focus on social amelioration increases as citizens’ interest in economic conditions increases, but only up to a point. As citizen's interest in economic conditions heightens further, the focus on social amelioration diminishes as the government shifts its emphasis on vaccination to convince the public to have themselves inoculated.

Results

We begin with a descriptive and visual examination of the data and its component variables. A logistic regression entails modeling the probability of the presence or existence of a particular event or class. In this case, our interest is in the probability of COVID-19-related vaccination getting more attention than social amelioration as a thematic focus of government policy narrative. An event is construed operationally as one in which vaccination as a topic model in the press release corpus exceeds social amelioration as a topic share for a given date. The graph in Figure 2 gives an informal comparison between vaccination and social amelioration as topics in the aggregated press releases issued between February 2020 and May 2021. To facilitate visual assessment, the values are normalized for both variables of interest.

Figure 2.

COVID-19 vaccination and social amelioration as topic models. Note: Values are normalized to facilitate visual comparison.

As the graph suggests, the government's economic response to COVID-19 has been considerably oriented towards mitigating the social cost of the pandemic. However, the trend appears to dissipate and over time the government's policy position seems to shift towards mass vaccination. Interest in mass vaccination, on the other hand, had a steadily trending increase that became more conspicuous sometime in the middle of January. The vertical line corresponding to February 28, 2021 indicates the date when the Philippines officially received its first delivery of COVID-19 vaccines. The press touted the event as kick-starting the government's mass inoculation campaign (Tomacruz, 2020). The highest observed search interest was on March 1, 2021. Meanwhile, the highest reported search interest on social amelioration was on April 27, 2020. The variable pertaining to the government's shifting policy narratives takes only two values (1 and 0). In the dataset, more than a third (36 percent) of the observations refer to instances in which vaccination received more policy attention than social amelioration. Table 1 gives a descriptive summary of the variables.

Table 1.

Descriptive statistics.

Variable	N	Mean	S.D.	Min	Max
Shifting attention from social amelioration to vaccines	428	0.64	0.48	0	1
Interest in the economy	428	0.83	1.03	0.04	5.39
Interest in vaccines	428	1.50	1.36	0.03	8
Interest in mass testing	428	0.15	0.17	0	1.25
Interest in COVID-19 signs and symptoms	428	0.11	0.16	0	1
Interest in community lockdowns	428	1.19	3.92	0	59
Log number of new COVID-19 deaths	428	2.95	1.29	0	6.00
Log number of new COVID-19 cases	428	6.69	2.10	0	9.64
Vaccine as topic in policy narratives	428	0.10	0.04	0.05	0.25
Social amelioration as topic in policy narratives	428	0.10	0.02	0.05	0.17

In the models, a day lag is used for all independent variables to ensure causality.

Table 2 summarizes the estimates for three logistic regression models gauging the effect of citizens’ concern about the economy on the government's COVID-19 policy position. Model 1 is an unrestricted model containing both linear and squared specification for query interest on the economy, while controlling for other COVID-19-related search topics, as well as the number of daily confirmed cases and deaths. Model 2 specifies interest in the economy as having only a linear relationship but likewise includes the previously identified control variables. The hypothesis suggests that citizens’ growing concern about the economy would, at some point, cause the government's policy focus to pivot from social amelioration to mass vaccination. It is not enough for this assumption to have a theoretical basis, however; the data must also provide structural support. Hence, following Cohen (2013), we ran Box-Tidwell (Box and Tidwell, 1962) transformations for the models to check whether our data support a curvilinear conjecture. Box-Tidwell regressions for the first two models did not converge. Meanwhile, the Box-Tidwell transformation for a more parsimonious specification seems to suggest a curvilinear relationship for interest in the economy, interest in vaccines, and new confirmed COVID-19 cases.

Table 2.

Summary of logistic regression estimates.

	Model 1	Model 2	Model 3	Model 4
	Vaccine attention	Vaccine attention	Vaccine attention	Vaccine attention
Interest in the economy	1.412***	0.700***	2.084***	2.075***
	(0.367)	(0.160)	(0.451)	(0.455)
Interest in the economy squared	−0.208*		−0.368***	−0.365***
	(0.0899)		(0.109)	(0.110)
Interest in vaccines	−0.0631	−0.0425	0.222	0.274*
	(0.126)	(0.131)	(0.419)	(0.131)
Interest in vaccines squared			0.00969
			(0.0842)
Interest in COVID-19 testing	0.871	1.406
	(1.970)	(2.047)
Interest in COVID-19 symptoms	4.562*	4.586*
	(2.143)	(2.168)
Interest in community lockdowns	1.103**	1.103**
	(0.418)	(0.411)
New COVID-19 deaths (log)	−0.131	−0.109	−0.279	−0.273
	(0.147)	(0.149)	(0.146)	(0.146)
New COVID-19 cases (log)	−0.192	−0.183	0.856***	0.853***
	(0.105)	(0.104)	(0.222)	(0.220)
New COVID-19 cases (log) squared			−0.125***	−0.125***
			(0.0249)	(0.0248)
Vaccine availability	−2.706*	−2.922**	−2.564*	−2.581*
	(1.064)	(1.060)	(1.091)	(1.073)
Constant	0.573	0.611	0.516	0.514
	(0.456)	(0.448)	(0.461)	(0.460)
N	427	427	427	427
AIC	375.7	377.3	419.4	417.4
BIC	416.2	413.8	455.9	449.9

Standard errors in parentheses. ^*p < 0.05, **p < 0.01, ***p < 0.001.

To examine whether the squared terms have explanatory power even with streamlined parameters, we conducted link tests for model adequacy (Pregibon, 1980; Tukey, 1949). Similarly, the link tests seem to prefer the parsimonious parameters in Models 3 and 4. Model 3 includes squared terms for the three variables assumed to be curvilinear but vaccine interest in this specification does not appear to have a significant effect. Model 4, which restricts the assumption of a curvilinear relationship to interest in the economy and new confirmed cases, also suggests that the relationship between interest in vaccines and a shifting policy narrative is linear.

Both Models 3 and 4 support the hypothesis that citizens’ pocketbook assessments of the economy have both short-and long-term effects on the government's policy response. The curvilinear relationship suggests that as Filipinos anxiously inquire about their livelihood and the economy amidst the COVID-19 pandemic, the government responds by providing social safety nets as an initial response. But as such interest in their economic condition heightens due to prolonged economic disruptions, the government is constrained to shift its policy focus towards wide-scale inoculation. For a more tangible notion of how much citizens’ pocketbook assessment of the economy shapes government policy, we examine adjusted predicted probabilities at various magnitudes, while all other variables are set at their means. To facilitate interpretation, the graph in Figure 3 illustrates the relationship visually.

Figure 3.

Citizen's search interest in the economy and government's policy frame.

Note: The grey areas indicate confidence intervals.

In periods in which citizens’ apprehension about the economy is still manageable, policy attention on social amelioration is also low and government has room to explore other policy options such as mass vaccination in its narrative. As people become more anxious about worsening economic dislocations, however, policy attention to social amelioration also increases in an attempt to pander to citizens’ demand. As economic anxiety heightens further, policy attention shifts back to mass vaccination, possibly as a long-term response. In Model 4, the turning point is when interest in the economy is at 2.8. The mean of the economic interest index is 1.7. About 20 percent of the observations are equal or above the tipping point, suggesting that the curvilinear specification is justified.

Model 4 also suggests that there is only a linear relationship between citizens’ interest in vaccines and policy attention. The probability of the government highlighting mass inoculation in its policy narrative increases with the public's growing interest in vaccines. Under a linear specification, this relationship is straightforward and statistically very significant. Figure 4 depicts this relationship visually.

Figure 4.

Citizens’ search interest in vaccines and government's policy frame.

Note: The grey areas indicate confidence intervals.

As with an interest in economic conditions, the number of daily confirmed new COVID-19 cases has a curvilinear relationship with attention to vaccines in the government's policy narrative. Although no specific hypothesis was posited for this variable, the result is intuitive and consistent with the article's overall theoretical conjecture. Based on the findings, the government panders to citizens’ demand for ameliorative socioeconomic interventions as daily COVID-19 cases increase, but only up to a point. Figure 5 shows that as the number of daily infections increases, policy attention shifts to vaccines in the government's policy narrative.

Figure 5.

Daily new COVID-19 cases and policy frame.

Note: The grey areas indicate confidence intervals.

Conclusion

In this article, we show that the Philippine government's framing of its response to the pandemic has vacillated between easing socioeconomic dislocations through temporary social amelioration and convincing the public to get inoculated. As citizens become interested in the economic implications of the pandemic, the government responds by highlighting social amelioration in its policy narratives. However, as citizens’ search interest increases with mounting economic dislocations, the government is constrained to redirect the focus of its policy narratives towards mass vaccination.

In the media, this approach was heavily criticized for slowing down the rollout of vaccines and aggravating the economic stagnation brought by mobility curbs at the height of the pandemic (Gonzales, 2021). Securing and delivering vaccines to ensure immediate access to the most vulnerable people has long been a recognized public health policy (Carroll et al., 2015; Hasan et al., 2021), and the Philippines may have acted a little late in this regard. Some in the Filipino medical community blame delayed procurement for the slow pandemic recovery (Macaraeg, 2021), an indication of mismanagement according to some political coalitions (Ramos, 2021). Studies observed that when the second wave of infection hit the Philippines in 2021, the government was so preoccupied with reallocating resources to COVID-19 that it did not take long for the public health system to become overwhelmed (Uy et al., 2022). By April 2021, rising numbers of cases pushed many facilities to critical capacity, so much so that hospitals turning away both COVID-19 and non-COVID-19 patients had become commonplace (Morales and Lema, 2021).

Arguably, the patchy state of the healthcare system tells a lot about the quality of the government response. It may be contended further that the government's semblance of responsiveness was all part of performance politics in anticipation of the 2022 elections. These are valid claims worthy of further scholarly introspection. Their elaboration, however, is beyond the scope of this article. What seems to be certain is that, like its counterparts elsewhere, the Philippine government was motivated and constrained by whatever capacity and opportunity were available in the creation of its pandemic response (Capano et al., 2020; Rudan, 2021). If COVID-19 is straining the public health system of wealthy countries, conditions are worse in developing countries grappling with social protection and health financing (Shadmi et al., 2020). Even before the pandemic, the Philippine healthcare system was already ill-prepared with its poorly funded public hospitals and inadequate social health insurance (Obermann et al., 2006; Querri et al., 2018). Current missteps and poor strategy amidst the pandemic only magnified what is already obvious.

The government initially anchored its COVID-19 response on a social amelioration program intended to minimize the economic cost of business suspensions and lockdowns. Operating on the assumption that the pandemic will only be temporary, the government pandered to citizens’ demand for social assistance payments and the physical distribution of cash. But social assistance programming being expensive, inefficient, and susceptible to imperfect targeting (Gerard et al., 2020), public demand for additional economic assistance also became a wake-up call of sorts for the government to seek a long-term public health option. Policy learning without a doubt occurred late but nonetheless this is far from the conventional populist description of seeking easy solutions or spectacularizing the crisis response. This is also consistent with prior accounts of resourced-constrained governments with fragmented political institutions opting for incremental policy changes over comprehensive reforms when faced with a health crisis (Oliver, 2006).

Nevertheless, we do acknowledge methodological limitations whose exploration would no doubt provide a richer understanding of crisis policy framing in the Philippines. Television, for instance, remains the top news source for Filipinos and may provide another dimension to Filipinos’ information-seeking behavior if data can be derived meaningfully. Filipinos’ social media behavior also suggests a promising platform for analyzing the link between public opinion and public policy, subject to parameters that will minimize the noise and bias in generated data (Olteanu et al., 2019).

Our findings have a number of implications. While regimes matter in dissecting pandemic responses, there is a lot to learn from citizens’ emotions and their potential to catalyze desirable policy outcomes. Also, in developing countries with patchy healthcare systems, the difficulty balancing between economic and public health priorities could just as easily be a function of structural inadequacy as it is of administrative inefficiency. Finally, in future pandemic response it may be worthwhile for governments to keep in mind that policy learning is meaningless unless decision makers act with coherence, foresight, and a sense of immediacy.