TRANSITION REPORT 2014 Innovation in Transition

Empirical analysis

Bank credit and firm-level innovation: a first look Table 4.1 and Chart 4.3 (see p68) take a first look at the relationship between credit constraints and innovation. Firms are grouped into three categories: firms with loans (3,840 firms); firms without loans, but without any need for them (4,723 firms); and firms with no loans and an unfulfilled need for credit (2,762 firms). The third group contains all credit-constrained firms.

Looking at the likelihood of innovative activity, there is a striking difference between the firms with loans and the firms that are credit-constrained. Of the firms with an unmet need for credit, 11.0 per cent, 11.2 per cent and 8.8 per cent have engaged in product innovation, process innovation and R&D respectively over the past three years. When we look at the firms that have been granted loans, these percentages are significantly higher at 15.3, 16.6 and 14.2 per cent respectively. In other words, firms with loans are around 40 per cent more likely to innovate than those without access to credit.

Read more

A clear picture – albeit only a preliminary one – is beginning to emerge as regards the relationship between access to credit and innovative activity: firms that innovate tend to be those that apply for a loan and are granted one. Firms that do not demand a loan in the first place are the least likely to innovate, probably because their lack of interest in borrowing coincides with a lack of innovative capacity.

For those firms that have managed to obtain a bank loan – a third of all firms interviewed – Table 4.1 also contains information on the type of bank that lent to them. Only 17 per cent of firms borrowed from a state bank, 29 per cent from a private domestic bank and 54 per cent from a foreign bank. Is innovative activity affected by the type of bank that a firm borrows from?

Table 4.1 suggests not (as does unreported additional analysis). There is some evidence that the clients of state and foreign banks innovate more, but these differences are fairly small and statistically weak, and they disappear when controlling for other firm-level characteristics.

Chart 4.4 shows product and process innovation among credit-constrained and unconstrained firms in selected transition countries. In almost all countries unconstrained firms innovate more than credit-constrained firms.

Chart 4.5 plots data for the same set of countries. Here, the horizontal axis measures the percentage of firms that are credit-constrained, while the vertical axis indicates the difference between the innovative activities of unconstrained and constrained firms. That difference is a rough indicator of the aggregate sensitivity of innovation to firms’ credit constraints in any given country. It indicates the extent to which reducing credit constraints could boost firm-level innovation, given the current economic, political and institutional framework in the country.

The chart shows that in some countries (such as Azerbaijan and Ukraine) credit constraints remain rife among firms. However, in these countries there is also little difference between credit-constrained and unconstrained firms in terms of their innovative behaviour. Consequently, it may be that access to credit has little impact on innovation in these countries, with other constraints – such as an inadequately educated workforce or corruption (see Chapter 3) – having more effect. In contrast, in countries such as Belarus, Lithuania, Romania and Russia, not only are there large numbers of credit-constrained firms but access to bank loans seems to have a relatively large impact in terms of unleashing innovation.

Chart 4.6 indicates that even within countries, at town or city level, there is a strong negative correlation between firms’ credit constraints and innovative activity. The remainder of this chapter looks at this relationship in more detail.

There are two main reasons for conducting this additional analysis. First, a more rigorous investigation is needed to control for other firm-level characteristics, so that the “everything else being equal” condition holds as much as possible. Second, the strong negative correlation between credit constraints and innovation does not necessarily indicate that credit constraints cause a decline in innovation. It could be that causation runs the other way – that is to say, it may be that when firms innovate successfully, banks are more amenable to financing them, thereby reducing credit constraints. One way to address this concern is to consider only credit constraints that are driven by external non-firm-specific factors. To this end, the remainder of this chapter focuses on the impact of credit constraints stemming from exogenous variation in the local banking landscape that surrounds each BEEPS firm.


Access to bank credit is associated with increases in firm-level innovation
  Percentage of firms engaged in  
  Product innovation Process innovation R&D Observations
  (1) (2) (3)  

Firms with loans





Private domestic bank





State bank





Foreign bank





Firms without loans





No demand















Source: BEEPS V.
Note: This table reports univariate results on the relationship between access to bank credit and firm-level innovation. *, ** and *** indicate significance at the 10%, 5% and 1% levels respectively for a two-sample t-test for a difference in means with unequal variances. The t-tests compare all firms with loans (top row) with all credit-constrained firms (penultimate row).


Source: BEEPS V.
Note: “No credit needs” denotes firms with no need for bank credit. “Unfulfilled credit needs” denotes firms with a need for bank credit that have either decided not to apply or were rejected when they applied. “Fulfilled credit needs” denotes firms with a need for credit that have received a loan from a bank.


Source: BEEPS V.
Note: Unconstrained firms need loans and are able to borrow from banks. Constrained firms need loans, but either decide not to apply for one or are rejected by the bank when they apply.


Source: BEEPS V.
Note: The vertical axis measures the difference between the average core innovation indices (calculated as the sum of product and process innovation) for unconstrained and constrained firms.


Source: BEEPS V.
Note: Each dot represents a town or city that contains more than 10 BEEPS firms. The x-axis measures the percentage of credit-constrained firms, while the y-axis measures the average core innovation index, which is constructed by regressing the average core innovation observed in the relevant area on the percentage of credit-constrained firms and country fixed effects. The predicted values are then plotted on the y-axis.