The Impact of Earnings Announcements Before and After Regular Market Hours on Asset Price Dynamics in the Fintech Era

Abstract

With the recent increase in retail investor participation led by commission-less fintech trading applications and new features like fractional trading, we now have higher volatility and significantly quicker price changes. This makes it hard to make informed trading decisions. Moreover, these effects are exacerbated even further around earnings announcements days. In this paper, we use Nasdaq data feed at a minute frequency and show that there is a significant increase in the slope of the price–volume structure during extended hours (after-hours, or pre-market hours) as compared with the ones observed during regular market times. Our analysis shows that the liquidity is much less during the extended market hours. As such, earnings announcements of stocks during these times have a significantly larger price impact than those stocks that have their earnings announced during regular trading hours. This significant difference can be explained by observing the limit order book structures during these different trading periods. We suggest that the earnings announcements should not be made during extended hours given the significantly lower liquidity and thus, the significantly larger price impact that not only determines the prices for the next trading session but also sets the new “fundamental” price signals for the stocks.

1. Introduction

The most important features of stock markets are liquidity and fundamental value signaling. Liquidity implies that there is enough volume available at, or close to, the best quote, i.e., that there is a minimum price impact when investors trade. Fundamental value signal allows market participants to make informed trading based on the price formation process developed by market participants (Madhavan, 2000). Order books (market depth) data measure the supply and demand for liquid, tradeable assets. In measuring real-time supply and demand, market depth is used by traders to assess the likely direction of an asset’s price. The limit order book is resilient if it promptly reverts to its normal shape after large trades (Large, 2007).

With the rise in fintech, increased low-latency activity improves traditional market quality measures, decreasing spreads, increasing displayed depth in the limit order book, and lowering short-term volatility (Hasbrouck & Saar, 2013; Riordan & Storkenmaier, 2012). However, with the increased overall activity in the stock markets and increased retail investors’ participation, the market structure might have changed.

There is an increase in execution costs and volatility when limit order books are displayed publicly. An increase in transparency reduces liquidity (Madhavan et al., 2005). Roşu (2019) shows that competition among speculators reveals much private information to the public, and the value of the information decays very fast. This speed increases the aggressiveness of the traders, leading to higher price volatility. Malceniece et al. (2019), using the European equity markets data, show that high-frequency trading (HFT) leads to increased liquidity, which is somewhat offset by an increase in comovement. Comovement affects how shocks are transmitted between stocks and, thus, the level of systemic risk. HFT has two opposing effects on the cost of capital. Firstly, HFT leads to increased comovement in returns, and liquidity is associated with cross-sectional returns that increase the cost of capital. On the other hand, HFT also increases the level of liquidity, which tends to lower the cost of capital.

Eaton et al. (2022) contrast outages at Robinhood, a platform more popular with inexperienced investors than traditional retail brokers, showing that herding by inexperienced investors can create inventory risks and harm liquidity in stocks with high retail interest, while other retail trading improves market quality. Leverage attracts speculative traders who do not necessarily contribute to market quality. By providing leverage, financial intermediaries exacerbate speculation, which increases trading volume and their intermediation revenues, but this is at the expense of traders’ profits and social welfare (Heimer & Simsek, 2019). Retail trading activity positively affects the volatility of stock returns, which suggests that retail investors behave as noise traders (Foucault et al., 2011). Furthermore, Foucault et al. (2005) find that the proportion of patient traders in the population and the order arrival rate are the key determinants of the limit order book dynamics. Traders submit aggressive limit orders (improve upon quoted spreads by large amounts) when the proportion of patient traders is large or when the order arrival rate is low.

For this reason, markets with a high proportion of patient traders or a small order arrival rate are more resilient. Also, a reduction in the tick size reduces market resiliency and, in some cases, increases the average spread. Baker and Stein (2004) discuss that market liquidity can be a sentiment indicator. An unusually liquid market is one in which pricing is dominated by irrational investors, who tend to underreact to the information embodied in either order flow or equity issues. Thus, high liquidity is a sign that the sentiment of these irrational investors is positive and that expected returns are, therefore, abnormally low.

Market liquidity is an important indicator representing the market depth and showing the absorption power of risk premium on trading execution. It also affects price discovery and market efficiency. We can improve market efficiency by increasing market liquidity. More precisely, a decline in the market prices’ uncertainties and increased market liquidity leads to more efficient markets. Furthermore market liquidity is essential for financial system stability. Maintaining sufficient liquidity under normal conditions will autonomously improve market stability by expanding the market boundaries and improving the participants’ confidence in market sustainability. (Muranaga & Shimizu, 1999).

In recent times, with the increase in retail investors’ participation led by multiple-fintech based trading applications which make stock market trading easily accessible, along with features like fractional trading1 that make high-price stocks accessible to even lower income decile investors, we have observed significantly quicker price changes and higher volatility in the stock markets.

These changes could be further exacerbated around the earnings announcements day. In this paper, we study the impact of earnings announcements before and after regular market hours on the asset price dynamics.

To briefly explain the way market hours are set in stock exchanges, we present in Figure 1 the timeline of the trading hours during a day “T”. These hours vary across exchanges, as some allow pre-market hours from 4:00 to 09:30 as well.

Companies announce their earnings quarterly, providing the company’s profitability and other performance metrics. When companies release their earnings reports, they have a significant implication in terms of media attention, and, in turn, investors’ attention and stock prices. The earnings announcements influence the share prices, which will move up (down) depending on whether the previous quarter’s performance was good (bad). These changes in prices are usually fundamentally driven and are an important signal for investors to decide their investments. The influence of earnings announcements on stock prices is exacerbated if they are made during periods of low liquidity in the markets (for example, during the after-hours or pre-market periods). If the earnings are good in the previous quarter, the firms would most likely make their reports public early in the regular market hours for maximum media attention, and if it is bad, they would like to make the report public in the after-hours or pre-market hours to draw less media attention (Noh et al., 2021).2

Market liquidity may have been further negatively impacted by the recent rise in retail trading (helped by fintech technology), access to fractional trading, and limit orders, which have the potential to modify the behavior of prices and volumes in limit order books3 (Tripathi & Rengifo, 2024), making them steeper by either keeping the same volume as before and increasing the prices or by keeping the same prices and reducing the offered volume (in either or both the buy and sell sides). During the extended hours, these effects could be further exacerbated. The limit order books can be very steep, and the price impact functions may also be much larger during these extended trading hours. The higher steepness of the limit order books during these extended hours not only has significant higher impacts on the current price levels, but it can also potentially and significantly influence the price levels observed after the earnings announcements. In other words, given the low liquidity observed during these periods, good (or bad) news can have an amplified impact that may be larger than what could have been observed if the announcement were made during regular trading times.

These price level changes are not necessarily driven by fundamentals but by the limit order books’ liquidity constraints present in the after hours. As such, movements of these types do not contribute to the company’s fundamental valuation, which is expected to happen. This increase in demand when earnings are announced, paired with the low liquidity provided in the limit order books, could potentially have a domino effect in the dynamics of price formation by influencing regular professional investors by either making them reluctant to go to the market (and thus reducing their activity as liquidity providers or takers) or making them behave more like “gamblers” than professional investor (since there is a potential for significant gains). Figure 2 illustrates three potential shapes of the limit order books.

This figure illustrates three hypothetical limit order book structures. The ‘lob_regular’ line presents the structure during regular market hours. During extended hours (after-hours and pre-market) on the earnings announcements day, as the market expects good (or bad) news, they could adjust their expectations by increasing prices and keeping their supplies, i.e., the limit order book structure, which becomes steeper but keeps the same volume (‘lob_extended_samevol’). Another situation that can arise is that the limit order book becomes even steeper with a reduced volume (‘lob_extended_reducedvol’). These situations create abnormal price formation mechanisms that could translate into price volatility increases (as recently observed in the market) and, at the same time, that could influence professional investors’ behavior by making them reluctant to go to the market, contributing in this way to an even stronger supply reduction.

To better understand these aspects and the impact of earnings announcements before and after the regular market hours on asset price dynamics in the fintech era, we analyzed Nasdaq data at the minute level to test hypotheses that are a feature of efficient markets and relevant to aspects such as market liquidity and financial stability. We study how the order book dynamics are during the regular market hours vs. the extended market hours around the earnings announcements days. The main testable hypotheses of the paper are as follows.

H1. 

Order book depth decreases significantly in the extended hours around earnings announcement as liquidity providers withdraw.

H2. 

The price impacts are significantly higher during extended market hours around earnings announcements days due to lower trading volumes and higher spreads.

H3. 

The bid–ask spreads are significantly higher in the extended market hours compared to regular market hours around earnings announcements days due to lower liquidity.

Our baseline hypothesis is that the limit order books should be similar during pre-and post- earnings announcements, i.e., to buy/sell a certain volume of shares, the change in the price levels should be similar in the pre-and post- earnings announcement periods.

Based on our analysis in this paper, we find that the price impact factors are significantly higher due to the fact that the market liquidity is significantly lower and the spreads are significant higher in the after-hours and the pre-market hours. As such, we propose that the earnings announcements should be better made public during regular market hours to have a better and more transparent price discover mechanism.

Our paper contributes to the growing literature on fintech and market microstructure. This paper enhances our understanding of the impact of earnings announcement before or after the regular market hours on asset price dynamics in the fintech era. To our knowledge, this is one of the earliest papers to study the impact of earnings announcements during regular vs. extended market hours in the times of fintech and financial markets tools like fractional trading.

The rest of the paper is organized as follows. Section 2 describes the data. Section 3 presents a descriptive analysis of the price–volume relationships on the buy and sell sides during the regular hours in day T, after hours on day T, pre-market hours on day T + 1, and regular market hours on day T + 1. Next, in Section 4, we present the analysis by the top 5 and top 10 best quotes on the buy and sell sides, followed by Section 5 with analysis determined by fixed cumulative volume, discussing the implications for the stock markets. Finally, Section 6 concludes and presents policy recommendations.

2. Data

This section presents the details of the data sources used in the study. Limit order books (LOBs) are the primary source of market data. They provide real-time updates throughout the day, reflecting the trading dynamics. The LOBs reveal the market depth by listing the number of shares being bid on or offered at each price point. Market depth is the key indicator of liquidity and the potential price impact of sizable market orders.

We use Historical Total View-ITCH data from Nasdaq, providing historical tick-by-tick Nasdaq order data. It allows subscribers to track individual orders for equity instruments from placement to execution or cancelation. The data comprises time-sequenced messages describing the history of trade and book activity. Each message is time-stamped to the nanosecond and hence provides a detailed picture of the trading process and the state of the Nasdaq book.

Our data include orders present during the regular market hours (09:30–16:00) on day T and day T + 1, after-hours (16:00–20:00) on the trading day T, and pre-market hours (04:00–08:00) on day after earnings announcement (T + 1). We analyze all the quarters in 2020 and 2021. The stocks chosen for this study are Apple (AAPL), Amazon (AMZN), Google (GOOG), and Tesla (TSLA), as these are the stocks with the most traded value during the study periods. These stocks cover, on average, 15–25% of the daily traded value (see Figure 3).4 The dates for earnings announcements for the stocks under study are presented in

Table 1. Earnings announcement dates for APPL, AMZN, GOOG, and TSLA in 2020 and 2021.

Earnings Announcements Dates
Year	Quarter	AAPL	AMZN	GOOG	TSLA
2020	Q1	28 January	30 January	3 February	29 January
	Q2	30 April	30 April	28 April	29 April
	Q3	30 July	30 July	30 July	22 July
	Q4	29 October	29 October	29 October	21 October
2021	Q1	27 January	2 February	2 February	27 January
	Q2	28 April	29 April	27 April	26 April
	Q3	27 July	29 July	27 July	26 July
	Q4	28 October	28 October	26 October	20 October

This table lists the earnings announcement dates for each quarter in 2020 and 2021 for Apple, Amazon, Google, and Tesla.

.

This figure presents the average percentage of traded value for Apple, Amazon, Google, and Tesla during the trading day T for the months under study. We see that these four stocks account for 15–25% of the average daily traded value.

3. Price–Volume Analysis

In this section, we analyze the price–volume relationship on the buy and sell sides during regular hours on day T, after-hours on day T, and pre-market hours of the day after the announcements day (i.e., day T + 1). We plot the fifth best quote (buy/sell limit orders) at each minute during the trading hours and calculate the cumulative volumes offered up to the fifth best quote. Figure 4 presents the price–volume analysis for the buy side of Apple.5

These figures present the fifth best quote at each minute and the cumulative volumes to reach that price quote, while the buy/sell limit orders were placed for Apple in 2020Q1 on the buy side during the regular hours on the earnings announcements day (day T), after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1.

Note: Reg(T): regular market hours on the earnings announcements date T; AH(T): post-market hours on day T; Pre(T + 1): pre-market hours day T + 1; Reg(T + 1): regular market hours on day T + 1.

From this figure, we can see that during regular market hours of the announcements day (T) of 2020Q16, for Apple’s buy side, the average start and end price quotes were USD 311.85 and USD 316.51, respectively (the maximum and minimum price quotes in this period were USD 317.20 and USD 311.82, respectively). During the after-hours on day T, the average start and end quotes were USD 316.51 and USD 320.6, respectively (the maximum and minimum price quotes in this period were USD 326.40 and USD 314.89, respectively). Furthermore during the pre-market hours on day T + 1, we see that the average start and end quotes were USD 322.50 and USD 322.48, respectively (the maximum and minimum price quotes in this period were USD 322.50 and USD 322.42, respectively). Finally, during the regular hours on day T + 1, we see that the average start and end quotes were USD 324.90 and USD 325.51, respectively (the maximum and minimum price quotes in this period were USD 326.73 and USD 322.42, respectively). We observe that the end quote during regular market hours was USD 316.51, which jumped to USD 326.40 during extended market hours (after-hour) on day T, followed by USD 322.50 during the extended market hour (pre-market hour) on day T + 1, and then, on day T + 1, during regular market hours, the start quote was USD 324.90. This behavior clearly shows that the price setting mechanism that happened after hours has a direct impact on day T + 1 behavior. This result is consistent for all the stocks under analysis and across all the analyzed quarters.

Next, we look at the average cumulative volume traded (in thousands) until the fifth best quote was reached during these periods. We see that the average cumulative volumes at the start and end minute during the regular market hours on day T were 46,095 and 64,166 shares, respectively (the average cumulative volume traded during this period was 60,277 shares). During the extended market hours (after hours) on day T, the average cumulative volumes at the start and end minute were 64,166 and 46,974 shares, respectively (the average cumulative volume traded during this period was 43,493 shares, i.e., 27.85% lower than the regular volume average on day T). During the extended market hours (pre-market hours) on day T + 1, we see that the average cumulative volumes at the start and end minute were 27,148 and 46,032 shares, respectively (the average cumulative volume traded during this period was 34,278 shares, i.e., 43.13% lower than the regular volume average on day T). Finally, during the regular market hours on day T + 1, we see that the average cumulative volumes at the start and end minute were 60,961 and 75,589 shares, respectively (the average cumulative volume traded during this period was 61,275 shares, i.e., only 1.64% larger than the regular volume average on day T).

Once more, we can see that the average price quotes significantly increased during the extended market hours and stayed at the same increased level during regular market hours on day T + 1, while the average cumulative volume decreased significantly during the extended periods and reverted to the same levels as regular markets on day T + 1. In the example presented above, the price increase is linked to good news in the earnings announcements. The company posted USD 91.8 billion revenue, which was a 9% increase from the year-ago quarter and a record increase.7 In summary, lower supply and higher prices, facing a higher demand during after-hours, are the ones that set the prices that will be available for trading starting in regular hours at T + 1.8

In the price volume analysis, we consistently observe that volume traded is significantly lower in the extended market hours, which leads to decreasing market liquidity and market depth during this period. The consequence of this fact is that prices that will be used starting at regular times during day T + 1 will be based on prices set under lower volume/higher prices than those available during regular times at time T. As such, the good (or bad) news coming from the quarterly earnings announcements are not only “artificially” expanded but that also set the new prices available for regular trading hours starting on day T + 1.

4. Price Impact Factor Analysis Using the Top 5 and Top 10 Best Quotes

In this section, we perform the price impact factor analysis using the top 5 and top 10 best quotes. In this first part of the analysis, we fetch the 5 and 10 best quotes at each minute for the stocks under study. Then, we calculate the cumulative volume of shares of the stock till the 5 (or 10) “q₅” (“q₁₀”) best quote price is reached. We obtain the change in price, defined as $\mathsf{\Delta }{p}_5=p_5-p_1;\mathsf{\Delta }{p}_{10}=p_{10}-p_1$, and the change in cumulative volume is defined as $\mathsf{\Delta }{q}_5=q_5-q_1;\mathsf{\Delta }{q}_{10}=q_{10}-q_1$. Furthermore, we calculate the slope of the price–volume structure (i.e., price impact factor), ${PIF}_5={\displaystyle \frac{\mathsf{\Delta }{p}_5}{\mathsf{\Delta }{q}_5}};{PIF}_{10}={\displaystyle \frac{\mathsf{\Delta }{p}_{10}}{\mathsf{\Delta }{q}_{10}}}$.

4.1. Empirical Tests and Results—Buy and Sell Sides

We compute and compare the price impact factors (PIFs) at each minute on the earnings announcements dates during the regular market hours (day T), extended market hours (after hours) on day T, pre-market hours on day T + 1, and regular market hours on day T + 1. We perform the analysis for the buy and sell sides. In

Table 2. Price impact factor (buy and sell sides) using the best five best quotes at each minute.

PIF5	After-Hours (Day T) vs. Regular Hours (Day T)				Pre-Market Hours (Day T + 1) vs. Regular Hours (Day T)				Regular Hours (Day T + 1) vs. Regular Hours (Day T)
	AAPL	AMZN	GOOG	TSLA	AAPL	AMZN	GOOG	TSLA	AAPL	AMZN	GOOG	TSLA
Buy side
2020Q1	2.77	64.93	1.76	6.96	20.92	130.12	2.29	44.07	0.93	7.68	14.11	1.34
2020Q2	1.35	3.84	−1.36	5.11	23.45	209.17	1.34	24.68	0.98	2.64	0.48	1.63
2020Q3	2.47	1.22	6.90	4.28	13.92	116.02	86.89	42.41	0.54	0.96	0.95	5.55
2020Q4	-	4.13	51.11	1.48	-	45.03	45.11	4.94	-	2.15	1.73	0.44
2021Q1	5.19	1.18	4.23	1.14	5.71	60.31	8.86	2.39	1.17	2.35	1.00	0.52
2021Q2	1.33	-	1.26	0.82	8.45	-	2.38	3.67	1.38	-	0.20	0.34
2021Q3	1.42	0.84	-	3.16	6.97	80.79	422.04	15.28	0.79	3.11	2.83	2.44
2021Q4	-	-	2.55	1.63	-	-	121.15	4.12	-	-	1.10	-
Sell side
2020Q1	1.51	0.66	33.49	1.22	9.53	208.59	6.13	29.26	0.38	1.26	1.08	8.45
2020Q2	3.79	1.33	1.82	1.32	72.55	86.27	3.38	20.10	0.88	0.58	2.49	-
2020Q3	1.28	0.53	1.13	1.19	35.25	364.79	-	23.17	0.67	3.17	1.44	1.00
2020Q4	-	2.81	1.04	1.91	-	322.35	228.67	6.85	-	2.60	1.65	1.43
2021Q1	2.43	0.94	1.09	8.05	5.97	75.49	1.71	16.10	0.95	0.85	3.37	1.04
2021Q2	1.12	-	1.00	3.08	2.90	-	2.06	26.21	1.12	-	1.21	0.98
2021Q3	1.55	16.87	-	1.15	8.37	85.52	63.51	3.93	1.03	1.14	1.60	0.23
2021Q4	-	-	5.48	1.27	-	-	421.27	3.93	-	-	9.18	-

The table presents the price impact factors (slopes) using the five best quotes at each minute on the buy and sell sides during the regular hours on the earnings announcements day (day T), after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1. We perform the analysis for all the quarters in 2020 and 2021. We see that the averages of the slopes (buy side) during the regular hours on the earnings announcements day (day T) are significantly different when compared to after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1.

, we present the PIF comparison results based on the analysis of the five best quotes at each minute.9

Table 210 shows that the price impact factors during the regular hours on the earnings announcements day (day T) are significantly different when compared to after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1. We perform the analysis for all the quarters in 2020 and 2021.

For 2020 Q1, looking at the buy-side results, we see that the PIFs were higher in after-hour on day T when compared to PIF during regular market hours on day T (2.77, 64.93, 1.76, and 6.96 times higher); during pre-market hours on day T + 1, the PIFs were higher when compared to PIF during regular market hours on day T (20.92, 130.12, 2.29, and 44.07 times higher); and finally, during regular market hours on day T + 1, the PIFs were 0.93, 7.68, 14.11, and 1.34 times when compared to PIF during regular market hours on day T for Apple, Amazon, Google, and Tesla, respectively.11

A similar behavior is shown for the 2020 Q1 sell side. We see that the PIFs were significantly higher in the after-hour on day T when compared to PIF during regular market hours on day T (1.51, 0.66, 33.49, and 1.22 times higher); during pre-market hours on day T + 1 when compared to PIF during regular market hours on day T (9.53, 208.59, 6.13, and 29.26 times higher), and the regular market hours on day T + 1. The PIFs were 0.38, 1.26, 1.08, and 8.45 times when compared to PIF during regular market hours on day T, for Apple, Amazon, Google, and Tesla, respectively.

We can see that the PIFs are significantly higher in the extended market hours where the trading volume is lower, which leads to lower market liquidity in this period and decreased market depth. Thus, in the extended market hours, the price impacts are magnified and are driven by lower volume traded, rather than by stocks’ fundamentals. This result should be considered carefully as soon as the price dynamics during earnings announcements could potentially not be the best one that allows a correct price discovery mechanism that should be required by market participants, investors, lenders, and regulators alike.

4.2. Empirical Tests and Results—Bid–Ask Spread

Next, we compare the bid–ask spreads at each minute on the earnings announcements dates during regular market hours (day T), extended market hours, after-hours on day T, pre-market hours on day T + 1, and regular market hours on day T + 1. We perform the analysis for the buy and sell sides. We perform the analysis for all the quarters in 2020 and 2021.

Table 3. Bid–ask spread using the average of the five best quotes on buy and sell sides at each minute.

Average Bid–Ask Spread (Best Five)	After-Hours (Day T) vs. Regular Hours (Day T)				Pre-Market Hours (Day T + 1) vs. Regular Hours (Day T)				Regular Hours (Day T + 1) vs. Regular Hours (Day T)
	AAPL	AMZN	GOOG	TSLA	AAPL	AMZN	GOOG	TSLA	AAPL	AMZN	GOOG	TSLA
2020Q1	1.37	17.91	2.58	1.26	5.92	11.03	0.24	9.09	3.66	0.83	1.29	2.94
2020Q2	1.56	3.20	1.25	1.94	15.56	9.72	0.01	11.01	15.83	5.77	0.56	13.61
2020Q3	1.85	0.85	1.32	2.72	2.44	3.70	-	2.28	1.37	0.79	2.16	3.13
2020Q4	-	2.38	4.31	1.34	-	5.46	0.07	2.77	-	1.59	1.82	2.18
2021Q1	1.50	1.40	2.59	4.05	4.75	9.60	−0.04	5.78	1.08	2.25	1.34	3.21
2021Q2	2.25	-	4.13	1.47	23.67	-	0.24	8.90	11.83	-	2.03	0.28
2021Q3	0.95	12.78	-	1.84	4.77	6.14	-	3.05	1.65	1.50	0.42	0.46
2021Q4	-	-	2.23	0.47	-	-	18.27	0.44	-	-	18.91	-

The table presents the bid–ask spread using the average of the five best quotes on buy and sell sides at each minute level during the regular hours on the earnings announcements day (day T), after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1. We perform the analysis for all the quarters in 2020 and 2021. We see that the average bid–ask spread increases significantly post the earnings announcements during the after hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1.

12 shows that the bid–ask spreads using the average of the five best quotes on buy and sell sides at each minute level during the regular hours on the earnings announcements day (day T), pre-market hours on day T + 1, and the regular hours on day T + 1.

We see that the bid–ask spread during the regular hours on the earnings announcements day (day T) are significantly different when compared to after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1.

For 2020 Q1, looking at the buy-side results, we see that the bid–ask spreads were higher in after-hour on day T when compared to bid–ask spreads during regular market hours on day T (1.37, 17.91, 2.58, and 1.26 times higher); during pre-market hours on day T + 1, the bid–ask spreads were higher when compared to bid–ask spreads during the regular market hours on day T (5.92, 11.03, 0.24, and 9.09 times higher); and finally, during regular market hours on day T + 1, the bid–ask spreads were 3.66, 0.83, 1.29, and 2.94 times higher when compared to bid–ask spreads during the regular market hours on day T for Apple, Amazon, Google, and Tesla, respectively.

Once more, we can see that, in most cases, the bid–ask spreads increased significantly in the extended market hours where the trading volume was significantly lower than the ones present during regular trading times.

5. Analysis by Fixed Cumulative Volume

As a robustness check, in this section, we set the cumulative volume at 50,000 shares and collect the corresponding price for each stock at each minute (see Figure 5). Then, we compute the change in price, $\mathsf{\Delta }{p}_{50000}$, i.e., the change in price to place a buy/sell limit order for 50,000 shares of the stock under study. Furthermore we calculate the slope of the price–volume structure, i.e., the price impact factor, $slop{e}_{50000}={\displaystyle \frac{{\mathsf{\Delta }p}_{50000}}{50000}}$. Using this impact factor, we test the changes observed in the price–volume structure before and after the earnings announcements.13

Finally, we also calculate the number of steps required to place a limit order to buy/sell 50,000 shares at each minute level and denote this variable as $step{s}_{50000}$. Furthermore, we use the number of steps to calculate the average tick size (defined as the minimum price movement of the stock) at each minute level. We define the average tick size as $averageticksize={\displaystyle \frac{\mathsf{\Delta }{p}_{50000}}{step{s}_{50000}}}$.

This figure illustrates a hypothetical limit order book structure with a fixed cumulative volume of 50,000 shares.

5.1. Empirical Tests and Results—Price Impact Factors (Buy and Sell Sides)

Table 4. Price impact factor of the price–volume structure (buy and sell sides) while placing an order for 50,000 shares.

PIF50k	After-Hours (Day T) vs. Regular Hours (Day T)				Pre-Market Hours (Day T + 1) vs. Regular Hours (Day T)				Regular Hours (Day T + 1) vs. Regular Hours (Day T)
	AAPL	AMZN	GOOG	TSLA	AAPL	AMZN	GOOG	TSLA	AAPL	AMZN	GOOG	TSLA
Buy side
2020Q1	3.45	14.02	9.82	5.24	11.95	32.87	3.92	23.89	0.99	2.23	6.35	2.23
2020Q2	1.23	1.37	1.18	4.23	11.59	51.30	3.43	14.34	1.19	1.71	0.98	1.23
2020Q3	2.85	1.93	16.74	7.40	8.28	70.85	58.18	24.70	0.94	2.64	0.65	2.88
2020Q4	-	1.42	23.01	1.66	-	40.86	26.54	4.57	-	6.97	1.85	0.89
2021Q1	3.58	0.89	4.32	1.49	5.89	14.63	5.93	3.65	1.13	0.68	−0.85	0.55
2021Q2	1.42	-	7.69	1.40	6.01	-	4.93	3.65	1.48	-	0.86	2.27
2021Q3	1.43	1.21	-	3.33	5.70	19.74	146.91	8.59	1.15	3.11	1.53	3.85
2021Q4	-	-	0.92	2.63	-	-	28.07	1.69	-	-	0.45	-
Sell side
2020Q1	1.42	0.73	6.05	1.14	3.24	32.29	2.58	6.13	0.40	6.44	0.63	2.21
2020Q2	0.78	1.00	1.45	1.41	10.58	8.71	4.18	5.94	1.00	0.47	1.92	1.02
2020Q3	1.03	0.97	1.03	1.01	14.58	19.42	-	9.24	1.19	1.36	1.21	3.73
2020Q4	-	1.43	1.02	1.02	-	35.77	36.78	2.40	-	1.23	0.62	0.59
2021Q1	1.82	0.84	1.01	4.03	3.20	15.08	1.85	4.52	1.84	1.14	6.17	3.13
2021Q2	1.21	-	0.68	3.65	3.45	-	3.76	9.65	1.21	-	1.03	2.06
2021Q3	1.20	8.96	-	1.04	10.90	23.41	31.45	5.93	1.15	2.07	2.52	3.00
2021Q4	-	-	3.73	1.18	-	-	46.41	1.21	-	-	4.87	-

The table presents the price impact factors in placing an order for 50,000 shares on the buy and sell sides during the regular hours on the earnings announcements day (day T), after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1. We perform the analysis for all the quarters in 2020 and 2021. We see that the averages of the slopes (buy and sell sides) during the regular hours on the earnings announcements day (day T) are significantly different when compared to after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1.

shows that the averages of slopes or price impact factors, when we fix the cumulative volume at 50,000 shares during the regular hours on the earnings announcements day (day T), are significantly different when compared to after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1. We perform the analysis for all the quarters in 2020 and 2021.

Once more, we can see that, in most cases, the PIFs are significantly higher in the extended market hours, when the trading volume is lower, which leads to lower market liquidity in this period and decreased market depth. So, in the extended market hours, the price impacts are magnified and driven by lower trading volume rather than the fundamentals.

5.2. Empirical Tests and Results—Number of Steps (Buy and Sell Sides)

Next, we look at the number of steps.

Table 5. Number of steps needed to place an order for 50,000 shares pre-and post-FT (buy and sell sides).

steps50k	After-Hours (Day T) vs. Regular Hours (Day T)				Pre-Market Hours (Day T + 1) vs. Regular Hours (Day T)				Regular Hours (Day T + 1) vs. Regular Hours (Day T)
	AAPL	AMZN	GOOG	TSLA	AAPL	AMZN	GOOG	TSLA	AAPL	AMZN	GOOG	TSLA
Buy side
2020Q1	15	32	8	13	−20	4	10	10	−1	6	7	−3
2020Q2	0	−14	13	20	−33	−37	4	26	3	−12	5	−8
2020Q3	12	−31	36	29	7	−20	−39	20	−5	−26	−15	−5
2020Q4	-	23	20	23	47	−5	−55	20	-	−6	−7	0
2021Q1	14	−32	25	3	5	−12	−13	19	−1	−12	−3	2
2021Q2	8	-	35	3	36	-	−11	34	4	-	1	5
2021Q3	12	3	-	20	38	42	−38	24	1	24	−8	−3
2021Q4	-	-	15	6	-	-	−14	7	-	-	12	−53
Sell side
2020Q1	13	−8	16	5	−7	13	12	15	−2	−1	4	−6
2020Q2	−7	−21	11	1	−15	−23	−29	−20	−2	−5	−12	−7
2020Q3	7	−13	15	1	−4	−13	−53	8	−3	0	2	8
2020Q4	-	19	5	4	53	14	−45	25	-	20	2	4
2021Q1	23	−18	1	30	26	7	4	30	2	−2	−4	−1
2021Q2	6	-	−4	25	2	-	0	28	2	-	−3	11
2021Q3	8	16	-	1	4	−6	−65	26	2	−16	0	15
2021Q4	-	-	29	16	-	-	−28	−18	-	-	−12	−46

The table presents the number of steps to place an order for 50,000 shares on the buy and sell sides during the trading hours on the earnings announcements day (day T), post-market hours on day T, pre-market hours on day T + 1, and the regular market hours on day T + 1. We perform the analysis for all the quarters in 2020 and 2021. We see that the average number of steps (buy and sell sides) during the trading hours on the earnings announcements day (day T) are significantly different when compared to post-market hours on day T, pre-market hours on day T + 1, and the regular market hours on day T + 1.

shows the comparison of the average number of steps required to place a limit order for 50,000 shares during the regular hours on the earnings announcements day (day T) to after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1. We perform the analysis for all the quarters in 2020 and 2021.

In terms of the number of steps to complete an order to buy or sell 50,000 shares of a given stock, the results are not conclusive in the sense that they (the number of steps) increased during the extended market hours. This last result basically means that what matters with the price mechanism set by the limit order books is related to the (significantly) lower supply of shares as well as the (significant) price difference (tick size) between steps. Once more, this result has serious policy implications for market microstructures. In the next section, we analyze the tick size dynamics observed.

5.3. Empirical Test and Results—Average Tick Size (Buy and Sell Sides)

Table 6. Average tick size to place an order for 50,000 shares pre-and post-FT (buy and sell sides).

Average Ticksize50k	After-Hours (Day T) vs. Regular Hours (Day T)				Pre-Market Hours (Day T + 1) vs. Regular Hours (Day T)				Regular Hours (Day T + 1) vs. Regular Hours (Day T)
	AAPL	AMZN	GOOG	TSLA	AAPL	AMZN	GOOG	TSLA	AAPL	AMZN	GOOG	TSLA
Buy side
2020Q1	2.55	8.53	8.40	4.21	5.78	18.09	2.49	19.63	1.02	2.03	5.68	2.52
2020Q2	1.21	1.69	0.92	2.58	7.48	23.86	1.79	8.93	1.12	2.07	0.89	1.45
2020Q3	2.31	3.23	8.84	4.38	5.11	45.87	23.62	13.54	1.06	3.99	0.85	3.14
2020Q4	-	1.04	17.06	1.10	-	24.71	11.92	3.15	-	8.07	2.08	0.89
2021Q1	2.62	1.55	2.63	1.27	3.11	8.82	6.47	2.31	1.16	0.73	0.92	0.51
2021Q2	1.08	-	3.99	1.29	2.69	-	3.59	2.04	1.30	-	0.85	2.03
2021Q3	1.07	1.15	-	2.34	2.58	10.46	50.05	5.58	1.12	2.18	1.80	3.90
2021Q4	-	-	0.75	2.29	-	-	12.49	1.21	-	-	0.35	-
Sell side
2020Q1	1.09	0.89	4.59	1.02	1.85	20.76	1.24	4.01	0.43	7.22	0.58	2.50
2020Q2	0.95	1.46	1.23	1.23	4.87	3.64	3.20	3.48	1.06	0.53	2.36	1.21
2020Q3	0.89	1.27	0.82	0.96	6.57	8.49	-	4.65	1.31	1.39	1.19	3.43
2020Q4	-	1.04	0.95	0.95	-	17.88	18.20	1.55	-	0.97	0.61	0.55
2021Q1	1.13	1.25	0.99	2.43	1.83	7.52	1.63	2.63	1.72	1.30	6.62	3.19
2021Q2	1.01	-	0.73	2.22	1.54	-	2.99	5.06	1.14	-	1.10	1.72
2021Q3	0.97	7.05	-	1.00	4.98	14.49	10.22	2.91	1.07	2.72	2.46	2.35
2021Q4	-	-	2.26	0.86	-	-	19.30	0.80	-	-	5.97	-

The table presents the average tick size to place an order for 50,000 shares on the buy and sell sides during the regular hours on the earnings announcements day (day T), after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1. We perform the analysis for all the quarters in 2020 and 2021. We see that the averages of the slopes (buy and sell sides) during the regular hours on the earnings announcements day (day T) are significantly different when compared to after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1.

shows that the averages tick size, when we fix the cumulative volume at 50,000 shares during the regular hours on the earnings announcements day (day T), are significantly different when compared to after-hours on day T, pre-market hours on day T + 1, and the regular hours on day T + 1. We perform the analysis for all the quarters in 2020 and 2021.

We can see that, in most cases, the average tick sizes are significantly higher in the extended market hours, when the trading volume is lower, which leads to lower market liquidity in this period and decreased market depth. So, in the extended market hours, the price impacts are magnified and driven by lower trading volume, rather than the fundamentals.

6. Conclusions and Policy Discussion

In this paper, we study the impact of earnings announcements on asset price dynamics in the fintech era. We perform three types of analysis, which allow us to show that the market dynamics significantly change around earnings announcements during regular and extended market hours. We observe that the volume traded decreased significantly in the extended market hours, leading to lower market liquidity and decreased market depth. H1: The order book depth decreased significantly in the extended hours around earnings announcement as liquidity providers withdraw. Furthermore we find that (H2) the price impacts are significantly higher during extended market hours around earnings announcements days due to lower trading volumes and higher spreads. Furthermore, we find that (H3) the bid–ask spreads are significantly higher in the extended market hours compared to the regular market hours around earnings announcements days due to lower liquidity.

The results suggest that the price impact factors and tick sizes increase significantly while the volume decreased significantly around the earnings announcements dates during the extended market hours (after hours on the earnings announcements date and pre-market hours on the next day). This contributes to higher volatility and significant quick price changes in the market, not driven by economic factors but by market imperfections observed during the extended market hours. It is important to note that the price levels set during the extended market hours remain similar during the regular trading hours of day T + 1 and that these new prices are the ones that set the starting point for the price dynamics of the coming quarter. As such, the price levels available on T + 1 regular trading days do not necessarily reflect the fundamental value of a given corporation, making it hard for investors to make decisions while they trade. We suggest that the earnings announcements should not be made during the extended hours (after-hours or pre-market hours), given the significantly lower liquidity and other market imperfections during these periods.

As a next step, we are studying the impact of fintech and trading application features, such as fractional trading during regular trading days on liquidity and volatility during the regular vs. extended-markets hours. This will allow us to delve deeper into the market microstructure in the fintech era and understand the order book and market dynamics better.