Table 4 represents the descriptive statistics of variables before and during COVID-19. From Table 4, we can see that the mean and median values of total debts, short-term debts, long-term debts, total assets, total leverage, short-term

Table 4. Summary statistic.

Note: Totaldebts, Shortdebts, Longdebts, Totalassets are the values in the table multiplied by 10 billion. *** represents significance at the 1% level.

leverage, long-term leverage all present a significant increase while ROA experiences a great decrease during COVID-19 period. As mentioned earlier, China has implemented a deleveraging policy since 2015, and the leverage ratio has started to decline in 2017. Therefore, we can preliminarily draw a simple conclusion that the production and business activities of the enterprise have been badly disrupted, and the enterprise suffered severe loss of yield and have to increase the leverage to survive the crisis during the pandemic period. It is worth noting that before the epidemic, the mean value of total leverage (total debts) is 0.416 (0.999 billion yuan) and the short-term leverage (debts) is 0.326 (0.676 billion yuan), while the mean value of long-term leverage (debts) is only 0.089 (0.309 billion yuan). This indicates that short-term debts is the main component of the external borrowing of non-financial enterprises in China. Next we explore the structural characteristics of leverage growth from three perspectives of industry, firm size and profitability level.

3.2.1. Structural Characteristics in Industry

Figure 2 describes the average total leverage ratios in pre- and during-COVID period for all industries in the whole sample. Figure 3 depicts the mean values of leverage ratio¹ from the fourth quarter of 2016 to the fourth quarter of 2022 for all sample industries. From Figure 2 and Figure 3, we can obtain that the leverage ratio during the pandemic increased significantly for most industries. Only the real estate industry (K) and the scientific research and technology services industry (M) experienced almost no increase², while the comprehensive industry (S)

Figure 2. The average leverage ratios before and during COVID-19 for each industry.

Figure 3. Leverage ratios of 17 industries from 2016q4 to 2022q4.

saw a great decline in financial leverage. This indicates that the growth of leverage ratio for China’s non-financial enterprises exhibits clear industry heterogeneity during the pandemic period, which may be caused by the combined effects of the characteristics of the epidemic impact, internal decision-making of enterprises, and external policy intervention.

During the epidemic, the government introduced stable growth policies and increased investment in public services such as environmental protection, education, healthcare, and public facilities. Enterprises in environment and public facilities management (N), education (P) and health and social work (Q) industry were easier to obtain low-interest loans, which encouraged them to proactively expand debt financing projects and showing an active increase in leverage ratio.

Business activities of companies in the industries of accommodation and catering (H), transportation, storage and postal services (G), and culture, sports and entertainment (R), as offline gathering industries, were restricted after the outbreak of COVID-19, resulting in a cliff-like drop in revenues, while rigid operating costs such as rent, personnel salaries, and equipment depreciation still had to be paid. These companies faced a serious cash flow shortage in maintaining basic operations. In response, the Chinese government introduced anti-epidemic credit support policies, such as prohibiting financial institutions from blindly withdrawing, cutting off, or suppressing loans, and allowing extensions or renewals for enterprises facing difficulty in repaying maturing loans. In addition, the leverage ratios of the three industries were below the warning line of 0.5, indicating a moderate-to-low level with controllable short-term debt pressure. As a result, enterprises in these industries had to rely on extending existing debts and/or borrowing new loans to survive the pandemic, leading to a passive significant increase in leverage ratios.

More than 70% of enterprises in the mining industry (B) and electricity, heat, gas and water production and supply (D) are large state-owned enterprises, and the leverage level in pre-COVID was below the warning line 0.6, possessing strong debt paying ability. More importantly, they have better access to gain bank loans based on their good credit and relationship with bank, making the leverage ratios stable with an upward trend during the pandemic.

As the industry with the highest leverage ratio, real estate companies (K) had long been in a state of high debt, and their development had entered a turning point before COVID. After the outbreak of the pandemic, national policies did not relax financing constraints on them. Instead, in August 2020, relevant regulatory authorities introduced the “three red lines” to directly limit the leverage of real estate companies. In December, “loan concentration” management was implemented, prompting real estate companies to shift from “ensuring cash flow” to “reducing leverage.” As a result, the real estate industry was actually in a state of deleveraging during the COVID period. However, as a highly leveraged sector, the construction industry (E) saw its leverage ratio continue to rise during the pandemic. The reasons may be as follows: on the one hand, output value and profits declined while epidemic prevention costs increased, leading to a rapid shrinkage of assets. On the other hand, operating cash flow income decreased, and in order to ensure the resumption of work and the continuation of strategic projects, enterprises actively issued new bonds or add new loans to maintain operations. More importantly, the expansion of special debt, such as special epidemic treasury bonds and local special debt, indirectly supported the financing availability of construction enterprises. Moreover, over 80% of construction companies are large enterprises that maintain good relationships with government departments. These factors resulted in a significant increase in the absolute value of debt.

Unlike the industries mentioned earlier, the industry of information transmission, software, and information technology services (I) was a sector that benefited from the COVID-19 pandemic and experienced a rapid decrease in leverage. During this period, demand for online office, online education, remote healthcare, e-commerce live streaming, and similar services surged, which directly boosted the industry’s business volume and revenue, thereby improving its cash flow. However, new infrastructure initiatives such as computing power, big data, and artificial intelligence require substantial upfront investment on research and development. For an industry with 73.3% private small and medium-sized firms, it was a good opportunity to expand market share and layout new tracks. Thus, it was often necessary to expand investment. Beyond their own funds, debt financing also serves as a vital source, which directly increased leverage.

Contrary to most industries, the comprehensive industry (S) is the only industry with a decrease in leverage ratio during the COVID period. Figure 4 describes the short-term and long-term leverage ratios and ROAs of the S industry from 2016q4 to 2022q4. As shown, the ROA declined in 2020q4, but sharply increased in 2021q4. Accordingly, the short-term leverage increased in 2020, but it immediately decreased in 2021. This may be related to the wide range of businesses involved, covering multiple industries. In this way, the industry has greater risk resilience, and is better able to withstand sudden exogenous shocks and quickly recovers profits.

Figure 4. Leverage ratios and ROA of comprehensive industry from 2016q4 to 2022q.

3.2.2. Structural Characteristics in Company Size

In this section, we analyze the structural characteristics of the capital structure of micro non-financial enterprises from the perspective of company size. We first calculate the mean value of each company’s size before the COVID period, then sort all sample companies in ascending order of company size and evenly divide them into four groups, and finally calculate the average leverage ratio of each group in each quarter. The smallest companies are classified into the 0% - 25% group, whereas the largest companies are classified into the 75% - 100% group. The results are reported in Figure 5.

As shown in Figure 5, the larger the company size, the higher the leverage ratio. Interestingly, a clear stratification pattern emerges: the difference in leverage ratios between the 75% - 100% and 50% - 75% groups exceeds 0.1, whereas the difference between the 50% - 75% and 25% - 50% groups is relatively smaller, slightly below 0.1. Moreover, the difference between the 0% - 25% and 25% - 50% groups narrows further to less than 0.06.

We can also observe that the financial leverage of the four groups of companies experiences significant growth throughout the entire sample period. It is worth noting that leverage ratios begin to increase at a faster pace after the outbreak of

Figure 5. Leverage ratios grouped by enterprise size from 2016q2 to 2022q4.

the COVID-19 pandemic, and the smaller the company size, the more rapidly its leverage ratio rises, which leads to a central tendency. That is to say, the gap between groups becomes increasingly narrower over time.

The enterprises in the top two groups were high-leverage companies³ before the epidemic, with weak long-term solvency and heavy debt repayment pressure. If they had taken on too much additional debt during the epidemic, they would have faced the risk of bankruptcy. In addition, in the top group, 60.63% of companies are state-owned firms, as shown in Table 5. High-leverage state-owned companies have been the key targets of the deleveraging policy since 2017. Therefore, the leverage of the top group barely grew during the COVID period. In the bottom two groups, over 64% are small and medium-sized private enterprises with relatively low leverage levels, and they are not subject to deleveraging pressure. Furthermore, the Chinese government has implemented a series of financial support policies to help small and medium-sized enterprises overcome difficulties caused by losses during the pandemic. Thus, enterprises in the bottom two groups

Table 5. Distribution of enterprises with different ownership structures (group by company size).

had to obtain more external loans to increase cash flow, and their leverage ratios exhibited significant growth.

3.2.3. Structural Characteristics in Corporate Profitability

In this section we explore the structural characteristics of the micro leverage ratio for non-financial enterprises from the perspective of ROA⁴. First, the average ROA of each company is calculated using a method similar to that in Section 3.2. All sample enterprises are then sorted in ascending order by ROA and equally divided into four groups, after which the leverage ratio is calculated for each group. Companies with the lowest ROA are assigned to the group of 0% - 25%, while those with the highest ROA are placed in the group of 75% - 100%. Figure 6 illustrates the leverage ratios grouped by ROA from 2016q2 to 2022q4.

From Figure 6, it is evident that the lower a group’s ROA, the higher its average leverage ratio, with noticeable differentiation across groups. The leverage ratios in the top two groups show a slight upward trend from 2016q2 to 2022q4, yet they remain nearly parallel, indicating that COVID-19 had almost no effect on their leverage. In contrast, the bottom two groups—the 0% - 25% and 25% - 50% groups—begin to diverge from the second quarter of 2016. Our focus is on the observation that the pace of differentiation accelerated more rapidly during the COVID-19 period than in the pre-COVID period. This suggests that companies with the lowest ROA⁵ and highest leverage in the bottom group experienced a rapid rise in leverage during the pandemic. Generally speaking, the weaker a firm’s profitability, the less capacity it has to actively increase leverage during the epidemic. The rise in leverage may be attributed to the following reasons: affected by the pandemic, their profits deteriorated sharply, compounded by excessive

Figure 6. Leverage ratios grouped by ROA from 2016q2 to 2022q4.

debt accumulated before the outbreak. For large enterprises, financing channels narrowed, maturing debts defaulted, and assets shrank, leading to a passive increase in leverage. For small and medium-sized enterprises, they had to rely on targeted credit support from the government and financial institutions to survive, which likewise contributed to a passive rise in loans and leverage during the COVID-19 period.