COVID-19: Fear of pandemic and short-term IPO performance.

This study analyzes the relationship between COVID-19 related fear and short-term IPO performance. Though the average market-adjusted initial return of IPOs in the year 2020 is higher than that of the last four decades, it decreases if fear of pandemic increases. The evidence is robust when we use matching firm-adjusted initial returns. Next, we analyze the persistence of performance after the IPO date. The results show that the performance of IPO firms is more sensitive to the fear of the pandemic than the performance of similar existing firms.

Introduction

The recent outbreak of Coronavirus (COVID-19) has infected almost 27.57 million people and caused 475,000 deaths in the US, at the time of writing. The increasing number of new cases and deaths related to COVID-19 has created a palpable fear and uncertainty among market participants, e.g., investors and analysts. A growing stream of literature analyzes the impact of the fear of infection and death related to the COVID-19 pandemic on global stock market performance, for example, fear and global stock market performances (Lyocsa et al., 2020; Salisu and Akanni, 2020), fear and commodity price returns (Salisu et al., 2020), death, panic, and the US equity market performance (Baig et al., 2020), the reactions of stock prices in the airline and tourism industry during the COVID-19 period (Carter et al., 2020), and the importance of social trust on firm performance during the pandemic (Mazumder, 2020), among many others. However, very few studies have analyzed the impact of fear of the pandemic on IPO underpricing and post-initial performance.1 Though the IPO underpricing puzzle has drawn considerable attention to both investors and researchers for decades (Bajo and Raimondo, 2017; Ibbotson, 1975, among others), the extent to how much investors’ fear of the pandemic affects underpricing begs further investigation.

The pandemic has resulted in a substantial economic meltdown throughout the world. In the United States, stock market volatility increased sharply in late February, and the stock market plunged almost 33% in one month from an all-time high level.2 However, although the market retreated in late March 2020, the market volatility remained well above the normal level. Given the uncertainty, it is surprising that the average initial return of IPO firms in 2020 is 9.30% higher than those of other years.3 Thus, it is appropriate to analyze the initial returns amid fear of the pandemic. In this study, we explore two relevant questions: first, whether IPO initial returns are sensitive to the overall fear and uncertainty due to the pandemic; second, whether the post-IPO date performance is more or less sensitive to fear of the pandemic.

Existing studies find that media coverage about IPO and the tone of reporting may also shape investors’ beliefs (Bajo and Raimondo, 2017). In this study, we argue that the fear associated with the COVID-19 pandemic creates remarkable uncertainty in the investment decision-making process. In one study, Bali et al. (2017) find that low uncertainty is associated with 6% higher annualized returns, as compared to the stocks of high uncertainty. Hence, we propose our first hypothesis, that underpricing is negatively associated with overall fear related to the pandemic. Next, we argue that fear affects IPO firms' stock performance more than that of existing firms. Uncertainty about IPO firms’ future growth opportunities is higher, and increases heterogeneity in investors’ beliefs that results in a higher crash risk (Hong and Stein, 2003). Moreover, information asymmetry is more pronounced for new firms because old firms have more available information (Dasgupta et al., 2010). Thus, we offer a subsequent hypothesis that fear more adversely affects IPO firms’ stock returns than those of existing firms.

To test the hypotheses, we use 81 IPO firms in the year of 2020. Following Salisu and Akanni (2020) and Salisu et al. (2020), we measure the fear index as an equally weighted index of both the daily cases and the daily deaths in the US (see Appendix A.2 for more details). In baseline regression, the results show that the average initial return (average market-adjusted initial return) decreases by 18.6% (17.70%) if the fear of pandemic prevails high in the market. The association is more robust when we use matching firm-adjusted returns adopting the propensity score matching (PSM) approach.4 The matching firm-adjusted initial return (Initial Delta Return) decreases by 23.9% (22.9%) if the fear index dummy increases from 0 to 1. The results are both statistically and economically significant. Upon testing the second hypothesis, we find that IPO firms are more affected by fear of the pandemic. We find that if the fear of pandemic increases by one standard deviation, IPO firms’ daily returns (Adj. daily returns) decrease by 0.34% (0.18%).

To our best knowledge, this study is the first to analyze first-day IPO performance during the COVID-19 crisis period. This study contributes to the prevailing literature as follows: first, the study contributes to current IPO underpricing literature from a new perspective that fear of the pandemic has explanatory power to explain IPO underpricing. Second, we compare post-IPO performance with that of existing firms. Thus, market reactions due to the pandemic fear for both IPO and existing firms will be an exciting addition to the current literature.

The remainder of the paper is organized as follows. Section 2 briefly reviews the existing literature and develops hypotheses. Section 3 describes the sample and data. Section 4 presents the empirical results. Section 5 concludes.

Literature review and hypothesis development

Empirical evidence of high first-day returns for IPO firms is well known as IPO underpricing. Ibbotson (1975) is the first who documents the high first-day returns. Numerous studies following this study document significant initial day return for IPO stocks and provide differential explanations for underpricing, such as information asymmetry between investors (Rock, 1986), the reputation of underwriters (Beatty and Ritter, 1986; Megginson and Weiss, 1991), signaling by qualitative firms (Grinblatt and Hwang, 1989; Welch, 1989), and so on. Several other firm-level attributes can explain IPO underpricing, such as ex-ante uncertainty of issuing firms (Beatty and Ritter, 1986; Betty & Zajac, 1994), uncertainty about future growth opportunity and firm age (Ritter, 1984; Loughran and Ritter, 2004), higher P/E ratio (Chen et al., 2004; Engelen, 2003), the proportion of insider shareholding (Habib and Ljungqvist, 2001). These explanations are based on the competitive theories, e.g., information asymmetry, signaling, market timing, agency theory, etc.

Existing studies predominantly explain the firm-specific factors that can explain the variation; however, they mostly ignore outside factors or exogenous shock that may affect IPO's initial day return. To mitigate the gap, Engelen and Van Essen (2010) document that country-level factors, such as Rule of Law, Anti self-dealing index, and corruption, can also explain almost 10% variation in the IPO underpricing. A growing stream of research advocates investor sentiment as a critical factor for IPO initial day return (Chen et al., 2020b; Derrien, 2005; Wang and Wu, 2015). Issuing firms time the market and more firms go public when investors’ sentiment is high (Lee, Shleifer, and Thaler, 1991) to coincide with periods of excessive market valuation (Baker and Wurgler, 2002). Zhao, Xiong, and Shen (2018) find that investor attention positively affects first-day underpricing. Media coverage (Bhattacharya et al., 2009; Chen et al., 2020a; Guldiken et al., 2017; Pollock and Rindova, 2003) and media tone (Bajo and Raimondo, 2017; Zou et al., 2020) for the IPO also influence the initial day return. Twitter sentiment can also explain the IPO underpricing, especially the pre IPO dates tweets (Liew and Wang, 2016). In one study, Loughran and McDonald (2013) claim that negative sentiment embedded in the S-1 forms is positively associated with underpricing.

Behavioral finance studies reveal that investors' negative sentiment and mood affect the decision-making and asset pricing. For example, Kaplanski and Levy (2010) find that the market loses more than $60 billion for each of the aviation disasters. Schmeling (2009) finds that consumer confidence as a proxy of investors sentiment can predict stock returns for 18 industrialized countries. This study concentrates on how a pure exogenous shock, COVID-19, affects the IPO initial day return during the pandemic period. The number of COVID-19-related cases and deaths has become an integral part of media coverage, which causes fear among investors and shapes their sentiment. Da et al. (2015) measure households' fear index based on textual search and find that fear is negatively associated with stock returns. According to Salisu and Akanni (2020), the fear sentiment is associated with a decline in stock price. Thus, we hypothesize the following:

The higher the fear index, the lower the initial return for IPO stocks.

We further expect that the impact of pandemic fear is more pronounced in the subsequent daily returns of the newly listed firms. IPO firms are typically young, immature, and relatively informationally opaque (Ljungqvist, 2007). Daspgupta, Gan, and Gao (2010) show that information asymmetry exists to a greater degree in newly listed firms because information availability is more for older firms. Kelly and Ljungqvist (2012) show that stock prices fall as asymmetry of information increases. Further, Dierkens (1991) shows that information asymmetry negatively affects stock price when firms announce a seasoned equity offering. Given the existence of more significant information asymmetry for newly listed firms and the fear adversely affecting investor sentiment, we hypothesize the following:

The higher the fear index, the lower the subsequent daily return for IPO stocks.

Data and sampling

Sample construction

In this section, we describe how we construct two samples for our empirical analysis. Our first sample consists of initial public offerings from January-2020 to July-2020. On December 31, 2019, China reported to the World Health Organization (WHO) a string of pneumonia-like cases in Wuhan. In the United States, the first COVID-19 case was confirmed on January 20, 2020. Even though there was no official pandemic declaration before March-2020, the fear related to the disease prevailed before March, especially during January and February. Thus, we added the IPOs of January and February in our study. Since the fear index has been calculated based on reported cases and deaths, the index value is zero if there is no reported case or death. Moreover, we include January and February in the sample to get enough variation in the explanatory variable, the fear index.

Consistent with the IPO-pricing literature, we exclude IPOs that have an offer price of less than 5, are in financial and utility sectors, are not traded in NYSE, NASDAQ, and AMEX, and are ADRs. This leaves us with a final sample of 81 firms for initial return analysis in baseline regression. We collect price and financial data from the COMPUSTAT, SEC EDGAR, and Professor Jay Ritter's website. To construct a proxy for market return, we use the return data of the S&P 500. Following convention, we calculate the initial return and adjusted initial return for each firm as follows;where, CP is the closing price on the first day of trading and OP is the offer price.where, Initial  Return is the first-day return and R is the market return on that day.

Our second sample consists of listed firms' daily returns from January-2020 to August-2020, collected from the COMPUSTAT daily database.5 Following asset pricing literature, we exclude firms if the price is less than $1, if it is an ADR, or if it is not traded in NYSE, NASDAQ, and AMEX. This gives us 358,593 (346,337) firm-day observations in panel regression (panel regression with entropy balancing approach). Daily return is the difference between today's price and yesterday's price, scaled by yesterday's price. Daily adjusted return is the difference between daily return and market return. For baseline (cross-sectional) and panel regression analyses, we control for the year 2019 firm-level characteristics.

Summary statistics

Table 1 presents the summary statistics of the initial and adjusted initial return for IPOs from January-2019 to July-2020.6 In Panel A, we show the distribution of initial and adjusted initial returns by year. Careful analysis shows that both initial and adjusted initial return distributions in 2020 lean further to the right than that in 2019. The mean adjusted initial return is 27.5% in 2020, as opposed to 13.7% in 2019. In Panel B, we show the mean of initial and adjusted initial returns by month. Panel C reports the mean initial and adjusted initial returns of 2020 IPO firms segregating into two samples: before and after February 15, 2020.

Table 1

Summary statistics: IPO initial return

This table presents the summary statistics for the initial return and adjusted initial return of initial public offerings from January-2019 to July-2020. The sample consists of 216 firms. Initial return is the ratio of the difference between the closing price on the first day of trading and the offer price divided by the offer price.A Adjusted initial return represents a market-adjusted return, which is the difference between the initial return and the S&P500 return on the same day. Panel A shows the return distribution of IPO firms by their year of IPO. Panel B shows the mean of initial and adjusted initial returns of IPO firms by their year and month of IPO. Panel C shows the mean of initial and adjusted initial returns of 2020 IPO firms by segregating them into two samples.

	Mean	Std. Dev	p5	p25	Median	p75	p95	N
Panel A: Summary Statistics of IPO returns
Year: 2019
Initial Return	0.138	0.295	-0.192	0.000	0.006	0.284	0.794	135
Adj. Initial Return	0.137	0.295	-0.190	-0.005	0.009	0.283	0.802	135
Year: 2020
Initial Return	0.273	0.464	-0.073	0.002	0.051	0.458	1.237	81
Adj. Initial Return	0.275	0.461	-0.082	0.005	0.056	0.432	1.235	81
Panel B: Monthly IPO mean returns
		Year: 2019				Year: 2020
Month		Initial Return	Adj_Initial Return	# IPOs	Month	Initial Return	Adj_Initial Return	# IPOs
January		0.000	0.001	1	January	0.387	0.397	7
February		-0.052	-0.051	15	February	0.099	0.098	14
March		0.129	0.128	7	March	-0.006	0.048	3
April		0.256	0.254	15	April	0.195	0.185	6
May		0.181	0.182	16	May	0.177	0.178	10
June		0.323	0.318	17	June	0.283	0.284	21
July		0.114	0.114	17	July	0.456	0.455	20
August		0.065	0.069	8
September		0.177	0.176	9
October		0.107	0.104	8
November		0.042	0.040	16
December		0.118	0.115	6
Panel C: IPO mean returns before and after February 15, 2020
		Initial Return	Adj_Initial Return	# IPOs	Month	Initial Return	Adj_Initial Return	# IPOs
Before February 15		0.223	0.226	17	After February 15	0..286	0.287	64

All returns are in decimal.

Table 2 Panel A provides descriptive statistics for the control variables used in the baseline regression.7 The mean log of proceeds is 18.88. The average asset size for 2020 IPO firms is 438.16 million, when the 50th-percentile value is 44.05 million. The mean offer price for 2020 IPO firms is $15.18.8 Around 78% of the IPO firms in our sample are traded in NASDAQ. The average age of our sample IPO firm is 4.93 years. The mean market capitalization of IPO firms is $1288.19 million. In Panel B, we show the descriptive statistics for the panel regression variables, separated by IPO and Non-IPO firms.9 Fig. 1 displays the bar chart of initial returns, adjusted initial returns, and fear index by month.

Table 2

Descriptive statistics

This table presents the descriptive statistics for variables used in the baseline and panel regression. Panel A presents the descriptive statistics for all independent variables used in baseline regression. OfferSize is the natural log of total proceeds of IPOs. AssetSize is the total asset size in millions. Price is the IPO offer price. NASDAQ, a dummy variable, takes 1 if the firm is traded in NASDAQ and 0 otherwise. Age is the issuer age from incorporation. Hi-tech is a dummy variable equal to 1, if a firm has a certain SIC code (see Loughran and Ritter, 2004), and 0 otherwise. Underwriter reputation is the lead underwriter's reputation collected from Professor Jay Ritter's website. Market capitalization is calculated based on post-IPO shares and first-day closing price. Price revision is the percentage increase in the final IPO offer price from the midpoint of the high and low prices in the initial filing. Syndicate size is the number of venture capital co-investors in IPO. Venture capital (VC), a dummy variable, takes 1 if VC backs the IPO, and 0 otherwise. CEO founder, a dummy variable, takes 1 if the CEO is also a founder, and 0 otherwise. Board independence is the fraction of independent directors on the board. R&D represents R&D expenditures scaled by total assets. Log volume is the log of the first-day trading volume. FearIndex is an equally weighted index of both new cases and death related to the COVID-19 pandemic. Panel B presents the descriptive statistics for the variables (Non-IPO and IPO firms) used in the panel regression. Daily return is the difference between today's and yesterday's price scaled by yesterday's price. Adj. daily return is the market-adjusted return where S&P500 is the market. Log (Asset) is the natural log of total assets. Leverage is the sum of long-term and short-term debt scaled by total assets. ROA represents net income scaled by total assets.

Panel A: Descriptive Statistics of the Baseline Sample
	Mean	Std. Dev	p5	p25	Median	p75	p95	N
OfferSize	18.880	1.092	16.520	18.380	19.110	19.370	20.660	81
AssetSize	438.160	1552.028	0.050	0.390	44.050	156.100	1602.300	81
Price	15.180	14.687	6.000	10.000	16.000	19.000	26.000	81
NASDAQ	0.780	0.441	0.000	1.000	1.000	1.000	1.000	81
Age	1.784	0.881	0.693	1.099	1.792	2.303	3.219	81
Hi-Tech	0.098	0.300	0.000	0.000	0.000	0.000	1.000	81
Underwriter Reputation	7.408	2.254	3.001	7.001	8.501	9.001	9.001	81
Market Capitalization	1288.190	2385.072	42.600	154.070	540.270	1078.550	5519.850	81
Price Revision	0.049	0.444	0.000	0.000	0.000	0.000	0.000	81
Syndicate Size	1.506	1.762	0.000	0.000	1.000	3.000	4.000	81
Venture Capital	0.469	0.502	0.000	0.000	0.000	1.000	1.000	81
CEO Founder	0.481	0.503	0.000	0.000	0.000	1.000	1.000	81
Board Independence	0.785	0.125	0.500	0.750	0.833	0.875	0.889	81
R&D	0.207	0.154	0.000	0.000	0.341	0.804	1.000	81
Log(Volume)	15.209	1.130	13.030	14.439	15.321	15.846	17.024	81
FearIndex	0.490	0.283	0.000	0.420	0.490	0.580	0.990	250
Panel B: Descriptive Statistics of the Panel sample
Non-IPO firms
	Mean	p5	p25	Median	p75	p95	N
Daily Return	0.001	-0.069	-0.019	0.000	0.020	0.074	352,878
Adj. Daily Return	0.001	-0.056	-0.017	-0.001	0.015	0.060	352,878
Log(Asset)	7.407	4.151	6.182	7.462	8.637	10.620	352,878
Leverage	0.305	0.000	0.112	0.284	0.439	0.697	352,878
Hi-Tech	0.175	0.000	0.000	0.000	0.000	1.000	352,878
R&D	0.062	0.000	0.000	0.006	0.062	0.312	352,878
ROA	-0.015	-0.443	-0.013	0.035	0.075	0.168	352,878
IPO Firms
Daily Return	0.002	-0.068	-0.013	0.000	0.013	0.078	5,793
Adj. Daily Return	0.000	-0.070	-0.019	-0.003	0.016	0.075	5,793
Log(Asset)	2.078	-3.411	-1.143	2.998	5.050	8.333	5,793
Leverage	0.320	0.000	0.000	0.222	0.597	1.026	5,793
Hi-Tech	0.098	0.000	0.000	0.000	0.000	1.000	5,793
R&D	0.207	0.000	0.000	0.341	0.804	1.000	5,793
ROA	-0.738	-1.694	-0.493	-0.103	0.000	0.064	5,793

Fig. 1

This figure displays the average initial return, the adjusted initial return of IPOs, and the fear index. The sample period ranges from January-2020 to July-2020. The sample consists of 81 IPO firms from the year 2020.

Empirical results

Fear and initial returns

In Table 3 , we regress the initial returns on the fear index and control variables by using the following model:Where, IR is the initial returns or adjusted initial returns of the IPOs. HighFear is a dummy variable 1 if the fear index is more than the median value at lag IPO day, and 0 otherwise. X is a vector of control variables in the year 2019. We control IPO offer size, asset size, offer price, NASDAQ dummy, age, Hi-Tech dummy, underwriter reputation, market capitalization, price revision, syndicate size, venture capital dummy, CEO founder dummy, board independence, R&D, and log volume (following Brav and Gompers, 1997; Chahine et al., 2020; Krishnan et al., 2011; Loughran and Ritter, 2004; Vong and Trigueiros, 2010; Zhou and Sadeghi, 2019). 10 d is the Fama and French 49 industry dummies and ε is the white noise when standard errors are heteroscedasticity-consistent robust.

Table 3

Baseline regression

This table presents regression results of COVID-19 fear on initial returns and adjusted initial returns of IPOs from January-2020 to July-2020. Initial return is the ratio of the difference between the closing price on the first day of trading and the offer price divided by the offer price. Adjusted initial return represents a market-adjusted return, where S&P500 is the proxy of the market. Initial delta return is the difference between the year 2020 IPO firms' initial return and that of matched 2019 IPO firms using the PSM approach. Adj. Initial delta return is the difference between the market-adjusted initial return of the year 2020 IPO firms and that of matched 2019 IPO firms using the PSM approach. HighFear is a dummy variable 1 if the FearIndex is more than the median value, and 0 otherwise. FearIndex is an equally weighted index of both new cases and deaths related to the COVID-19 pandemic. OfferSize is the natural log of total proceeds of IPOs. Log(Assets) is the natural log of total assets. Price is the IPO offer price. NASDAQ, a dummy variable, takes 1 if the firm is traded in NASDAQ and 0 otherwise. Age is the issuer age from incorporation. Hi-tech is a dummy variable equal to 1, if a firm has a certain SIC code (see Loughran and Ritter, 2004), and 0 otherwise. Underwriter reputation is the lead underwriter's reputation collected from Professor Jay Ritter's website. Market capitalization is calculated based on post-IPO shares and first-day closing price. Price revision is the percentage increase in the final IPO offer price from the midpoint of the high and low prices in the initial filing. Syndicate size is the number of venture capital co-investors in IPO. Venture capital (VC), a dummy variable, takes 1 if VC backs the IPO, and 0 otherwise. CEO founder, a dummy variable, takes 1 if the CEO is also a founder, and 0 otherwise. Board independence is the fraction of independent directors on the board. R&D represents R&D expenditures scaled by total assets. Log volume is the log of the first-day trading volume. Industries are defined as Fama–French 49 industries. Standard errors are heteroskedasticity-consistent standard errors. t-statistics are displayed in parentheses. Significance levels are denoted by *, **, ***, which correspond to the 10%, 5%, and 1% levels, respectively.

	Initial Return	Adj. Initial Return	Initial Delta Return	Adj. Initial Delta Return
VARIABLES	(1)	(2)	(3)	(4)
HighFeard − 1	-0.1861**	-0.1770**	-0.2388**	-0.2293*
	(-2.1964)	(-2.0818)	(-2.0017)	(-1.9396)
Offer Sizet − 1	-0.0493	-0.0614	-0.1685	-0.1789
	(-0.3850)	(-0.4818)	(-1.1551)	(-1.2378)
Log(Assets)t − 1	0.0224	0.0237	0.0444	0.0451
	(0.5864)	(0.6355)	(1.0050)	(1.0511)
Pricet − 1	0.0847***	0.0844***	0.0716**	0.0715**
	(3.1277)	(3.1625)	(2.1193)	(2.1348)
NASDAQt − 1	-0.3474**	-0.3404**	-0.4299**	-0.4233**
	(-2.2534)	(-2.1790)	(-2.2775)	(-2.2199)
Age	0.0590	0.0576	0.0254	0.0246
	(0.6981)	(0.6954)	(0.2038)	(0.1994)
Hi-Tech	0.0913	0.0521	0.6341*	0.5939*
	(0.4092)	(0.2354)	(1.9386)	(1.8272)
Underwriter Reputation	-0.0091	-0.0057	0.0119	0.0151
	(-0.3185)	(-0.1962)	(0.3016)	(0.3806)
Market Capitalization	0.0000	0.0000	0.0001	0.0001
	(0.1395)	(0.1578)	(1.5690)	(1.5916)
Price Revision	0.1223**	0.1330**	0.2434***	0.2549***
	(2.1414)	(2.3575)	(3.6099)	(3.8125)
Syndicate Size	-0.0506	-0.0474	-0.0825	-0.0803
	(-0.8855)	(-0.8456)	(-1.5160)	(-1.4995)
Venture Capital	0.0888	0.0898	0.1839	0.1830
	(0.7321)	(0.7506)	(0.8843)	(0.8877)
CEO Founder	-0.0738	-0.0781	-0.0932	-0.0973
	(-0.7306)	(-0.7841)	(-0.7825)	(-0.8233)
Board Independence	0.4112	0.4515	0.1309	0.1738
	(1.1501)	(1.2827)	(0.2597)	(0.3509)
R&D	-0.0046	-0.0046	-0.0037	-0.0036
	(-1.0761)	(-1.0758)	(-1.0110)	(-1.0212)
Log(Volume)	-0.0739	-0.0734	-0.1106	-0.1104
	(-0.6607)	(-0.6628)	(-0.9466)	(-0.9495)
Constant	1.1196	1.2776	4.0168**	4.1474**
	(0.8008)	(0.9311)	(2.2846)	(2.3832)
Observations	81	81	81	81
R-squared	0.6892	0.6904	0.6686	0.6690
Industry FE	YES	YES	YES	YES

Table 3 presents the baseline cross-sectional regression results of 81 IPO initial returns on the fear index. We consider four proxies for the initial returns. In columns 1 and 2 of Table 3, we use two different initial returns as dependent variables, following Eqs. (1) and (2). The results show that the HighFear dummy is negatively and significantly associated with initial returns and adjusted initial returns after we control for firm-level variables and industry fixed effects. The results are economically significant, meaning that the fear index's dummy increase from zero to one is associated with an 18.6% (17.70%) decrease of initial (adjusted initial) return. The result is consistent with the notion that fear sentiment negatively affects the initial IPO returns for the year 2020. Next, we match each of the 2020-IPO firms with similar 2019-IPO firms to construct peer adjusted initial returns using the propensity score matching technique.11 Column 3 reports peer adjusted initial returns, while column 4 reports peer and market-adjusted initial returns. We find robust evidence when we use the PSM adjusted returns in columns 3 and 4, meaning that the prevalence of high fear in the market is associated with a 23.9% (22.9%) decrease of peer adjusted initial (peer and market-adjusted initial) return.

Fear and post IPO performance

Next, we extend our analysis to examine how the fear of COVID-19 affects the subsequent performances of IPO firms. Our panel dataset consists of daily stock returns data for the new and existing firms from January 2020 to August 2020. In Table 4 , we regress the subsequent daily returns on the fear index and control variables by using the following model: Where, DailyReturn is the daily return or market-adjusted daily returns, FearIndex is the fear associated with the COVID-19 at lag day, NewFirm is a dummy variable if the firm is an IPO firm in year 2020, and X is a vector of control variables. We take the following controls: log asset size, leverage, Hi-Tech dummy, R&D, and ROA. d is the Fama and French 49 industry dummies and ε is the white noise when standard errors are clustered at the firm level.

Table 4

Panel regression

This table presents regression results of COVID-19 fear on subsequent day return of IPO and non-IPO firms from January-2020 to August-2020. Daily return is the difference between today's and yesterday's price scaled by yesterday's price. Adj. daily return is the daily market-adjusted return. FearIndex is an equally weighted index of both new cases and deaths related to the COVID-19 pandemic. NewFirm is a dummy variable equal to 1 if the firm's IPO year is 2020, and 0 otherwise. Log (Assest) is the natural log of total assets. Leverage is the sum of long-term and short-term debt scaled by total assets. Hi-Tech is a dummy variable equal to 1, if a firm has a certain SIC code (see Loughran and Ritter, 2004), and 0 otherwise. R&D represents R&D expenditures scaled by total assets. ROA represents net income scaled by total assets. Industries are defined as Fama–French 49 industries. Standard errors are clustered by firm-level. t-statistics are displayed in parentheses. Significance levels are denoted by *, **, ***, which correspond to the 10%, 5%, and 1% levels, respectively.

	Daily Return		Adj. Daily Return
VARIABLES	(1)	(2)	(3)	(4)
FearIndexd − 1	0.0018***	0.0018***	-0.0017***	-0.0017***
	(4.6139)	(4.5984)	(-4.4325)	(-4.4815)
NewFirm	0.0024	0.0024	-0.0012	-0.0012
	(0.9500)	(0.9644)	(-0.5311)	(-0.5112)
FearIndexd − 1*NewFirm	-0.0120***	-0.0120***	-0.0065*	-0.0065*
	(-2.8763)	(-2.8505)	(-1.6977)	(-1.6843)
Log(Assets)	-0.0006***	-0.0006***	-0.0006***	-0.0006***
	(-8.9369)	(-8.8354)	(-8.5997)	(-8.4494)
Leverage	0.0019***	0.0018***	0.0019***	0.0018***
	(3.8037)	(3.4253)	(3.7816)	(3.3549)
Hi-Tech	-0.0000	-0.0001	-0.0000	-0.0001
	(-0.1739)	(-0.2283)	(-0.0907)	(-0.2999)
R&D	0.0035**	0.0039**	0.0033**	0.0037**
	(2.4099)	(2.3182)	(2.3284)	(2.2823)
ROA	-0.0004	-0.0004	-0.0004	-0.0004
	(-0.5402)	(-0.5432)	(-0.5243)	(-0.5319)
Constant	0.0045***	0.0046***	0.0052***	0.0053***
	(7.2071)	(7.0492)	(8.3862)	(8.1501)
Observations	358,593	358,593	358,593	358,593
R-squared	0.0005	0.0006	0.0005	0.0006
Industry FE	NO	YES	NO	YES

Table 4 presents the regression results of daily returns on the fear index. We consider two alternatives for daily returns. In columns 1 and 2, the daily return is the difference between today's and yesterday's price, scaled by yesterday's price. The columns are different in industry fixed effect treatment. In columns 3 and 4, we use market-adjusted returns as dependent variables. The only difference between the columns is the industry fixed effect. Here, the interaction term, FearIndex*NewFirm, captures the impact of fear sentiment on new firms. In column 3, the coefficient for FearIndex is negative and statistically significant, which confirms that fear sentiment negatively affects non-IPO firms’ expected return. Importantly, our variable of interest, FearIndex*NewFirm is negative and statistically significant. The result is economically significant, meaning that a one standard deviation increases in the FearIndex (0.2836) is associated with a 0.18% decrease in expected return. Our result is identical in column 4. Overall the results confirm that if there is no fear, old firms perform better than new firms suggesting that information asymmetry is lower for older firms (Dasgupta et al., 2010). When the fear sentiment is non-zero, its impact is more pronounced on the newly listed firms, which is consistent with our hypothesis.

Robustness: matching through entropy balancing techniques

To better identify how fear impacts daily returns of IPO and Non-IPO firms, we utilize a robust multivariate matching technique, known as entropy balancing (Hainmueller, 2012). This method ensures proper covariate balancing between treated (IPO firms) and control (Non-IPO firms) samples by weighing observations such that the post-weighting means and variance of treated and control firms are similar for each matching dimension. We match all the control variables used in Table 4. Table 5 Panel A represents the mean and variance of covariates of IPO and Non-IPO firms after entropy balancing. The difference in means and variances of covariates are minimal and statistically insignificant, suggesting that proper entropy balancing was achieved. Using the balanced sample with a post-weighting total of 346,337 firm-day observations, we re-run Eq. (4). We expect that the regression coefficient is free from any biases because the distribution of both treated and control samples are identical. Table 5 Panel B presents the regression results. Our key variable of interest, the interaction effect, is negative and statistically significant across all specifications. The result strongly supports our hypothesis that the impact of fear sentiment is more pronounced on the newly listed firms’ subsequent performances.

Table 5

Panel regression: entropy balancing approach

This table presents regression results of COVID-19 fear on subsequent day return of IPO and non-IPO firms from January-2020 to August-2020 using the entropy balancing approach. Panel A presents means and variances of control variables for both IPO and non-IPO firms used in regression after entropy balancing. Panel B presents the regression results using the post-weighting treated and control observations that were subject to entropy balancing. Daily return is the difference between today's and yesterday's price scaled by yesterday's price. Adj. daily return is the daily market-adjusted return. FearIndex is an equally weighted index of both new cases and deaths related to the COVID-19 pandemic. NewFirm is a dummy variable equal to 1 if the firm's IPO year is 2020, and 0 otherwise. Log (Assets) is the natural log of total assets. Leverage is the sum of long-term and short-term debt scaled by total assets. Hi-Tech is a dummy variable equal to 1, if a firm has a certain SIC code (see Loughran and Ritter, 2004), and 0 otherwise. R&D represents R&D expenditures scaled by total assets. ROA represents net income scaled by total assets. Industries are defined as Fama–French 49 industries. Standard errors are clustered by firm-level. t-statistics are displayed in parentheses. Significance levels are denoted by *, **, ***, which correspond to the 10%, 5%, and 1% levels, respectively.

Panel A: Difference in covariates after entropy balancing
Covariates	IPO Firms		Non-IPO Firms		Std. Diff
	Mean	Variance	Mean	Variance
Log(Asset)	5.3068	4.1418	5.3090	4.1515	-0.0011
Leverage	0.3804	0.1097	0.3805	0.1098	-0.0002
Hi-Tech	0.0584	0.0550	0.0584	0.0550	0.0000
R&D	181.0366	505756.3100	180.8938	505376.1300	0.0002
ROA	-0.5383	1.1694	-0.5381	1.1690	-0.0002
Panel B: Fear and IPO performance after entropy balancing
	Daily Return		Adj. Daily Return
	(1)	(2)	(3)	(4)
FearIndexd − 1	0.0014	0.0013	-0.0023**	-0.0024**
	(1.1514)	(1.0389)	(-1.9614)	(-2.0217)
NewFirm	0.0132**	0.0119**	0.0084	0.0071
	(2.2478)	(1.9835)	(1.5961)	(1.2937)
FearIndexd − 1*NewFirm	-0.0272**	-0.0285**	-0.0198**	-0.0211**
	(-2.4088)	(-2.5547)	(-1.9833)	(-2.1172)
Log(Assets)	-0.0016***	-0.0018***	-0.0016***	-0.0018***
	(-2.9073)	(-3.6107)	(-2.8562)	(-3.7109)
Leverage	0.0016	0.0018	0.0017	0.0020
	(0.4319)	(0.4648)	(0.4415)	(0.5073)
Hi-Tech	0.0053**	0.0017	0.0055**	0.0015
	(2.1957)	(0.5213)	(2.3680)	(0.4972)
R&D	0.0000	0.0000	0.0000	0.0000
	(0.2961)	(0.8005)	(0.3321)	(0.8945)
ROA	0.0011	-0.0015	0.0010	-0.0014
	(0.5953)	(-0.5792)	(0.5814)	(-0.5635)
Constant	0.0109***	0.0097***	0.0115***	0.0105***
	(3.0216)	(2.6582)	(3.2216)	(3.0302)
Observations	346,337	346,337	346,337	346,337
R-squared	0.0030	0.0044	0.0027	0.0042
Industry FE	NO	YES	NO	YES

Conclusion

We investigate the impact of fear associated with the pandemic on initial IPO returns, motivated by the nearly 9.30% higher IPO initial returns in 2020 than in the previous 40 years. We evaluate whether initial returns are sensitive to fear of the pandemic given the outperformance of the initial return. Using the fear index, we find that the initial return is negatively associated with the fear of the pandemic. The results are robust when we match the IPO firms with the previous year's IPO firms. Moreover, for post-IPO initial returns, public fear and sentiment affect newer firms more than older firms, even after entropy balancing. While this study is a preliminary analysis of the impact of pandemic fear on short-term IPO performance, the long-term performance of IPOs launched during the pandemic is a subject for future research.

Declaration of Competing Interest

We declare that we have no material financial interests that relate to the research described in this paper.

Variable name	Description	Source
Adj. Daily Return	Adj. daily return is the daily market-adjusted return. The market is defined as the S&P 500.	COMPUSTAT (Daily)
Adj. Initial Return	Market return adjusted Initial Return. The market is defined as the S&P 500.	COMPUSTAT (Daily), SEC EDGAR
Adj. Initial Delta Return	Adj. Initial delta return is the difference between the market-adjusted initial return of the year 2020 IPO firms and that of matched 2019 IPO firms using the PSM approach.	COMPUSTAT (Daily)
Age	The natural log of 1 + the age (in years) of the issuer at the time of the offering, as computed from the firm's first incorporation to the date of the offering.	Jay Ritter's Web site
Board Independence	Board independence is the fraction of independent directors on the board.	SEC EDGAR
CEO Founder	CEO founder, a dummy variable, takes 1 if the CEO is also a founder, and 0 otherwise.	SEC EDGAR
Daily Return	Daily return is the difference between today's and yesterday's price scaled by yesterday's price.	COMPUSTAT (Daily)
FearIndex	FearIndex is an equally weighted index of both new cases and deaths related to the COVID-19 pandemic.	HealthData.Gov & Salisu and Akanni (2020)
Hi-Tech	A dummy variable that equals 1 if the IPO is a high-tech firm and zero otherwise. In line with Loughran and Ritter (2004), high-tech firms are those with SIC codes 3571, 3572, 3575, 3577, 3578 (computer hardware), 3661, 3663, 3669 (communications equipment), 3671, 3672, 3674, 3675, 3677, 3678, 3679 (electronics), 3812 (navigation equipment), 3823, 3825, 3826, 3827, 3829 (measuring and controlling devices), 3841, 3845 (medical instruments), 4812, 4813 (telephone equipment), 4899 (communications services), 7371, 7372, 7373, 7374, 7375, 7378, and 7379 (software).	COMPUSTAT (Annual)
HighFear	A dummy is equal to 1 if FearIndex is higher than median value, 0 otherwise	HealthData.Gov & Salisu and Akanni (2020)
Initial Delta Return	Initial delta return is the difference between the initial return of the year 2020 IPO firms and that of matched 2019 IPO firms using the PSM approach.	COMPUSTAT (Daily)
Initial Return	The ratio of the difference between the closing price on the first day of trading and the offer price divided by the offer price	COMPUSTAT (Daily), SEC EDGAR
Leverage	The ratio of the sum of the book value of long-term (dltt) and short-term debt (dlc) divided by the total assets.	COMPUSTAT (Annual)
Log(Volume)	Log volume is the log of the first-day trading volume	COMPUSTAT (Daily)
Market Capitalization	Market capitalization is calculated based on post-IPO shares and first-day closing price.	COMPUSTAT (Daily)
NASDAQ	A dummy variable equal to 1 if the firm is traded in NASDAQ and 0 otherwise.	COMPUSTAT (Daily)
NewFirm	A dummy variable equal to 1 if the firm's IPO year is in 2020 and 0 otherwise	COMPUSTAT (Daily)
OfferSize	Natural log of total proceeds collected through IPOs.	SEC EDGAR
Price	Offer Price of IPO	SEC EDGAR
Price Revision	The percentage increase in final IPO offer price from the midpoint of the high and low prices in the initial filing	SEC EDGAR
R&D	R&D expenditures (XRD) scaled by total assets (at)	COMPUSTAT (Annual)
ROA	Net income (NI) scaled by total assets (at)	COMPUSTAT (Annual)
Underwriter Reputation	The lead underwriter reputation score	Jay Ritter's Web site
Syndicate Size	Number of VC co-investors in the IPO	SEC EDGAR
Venture Capital	A dummy variable takes 1 if the IPO is backed by VC, and 0 otherwise	Jay Ritter's Web site