The data on innovations, which we use in this study, are coming from two different sources. The first one is the enterprise survey, conducted specially for this paper by S. Tsukhlo from the Institute of Economics of Transition (IET). The innovation questioner was sent by mail to the sample of 1200 firms, which usually participate in the monthly surveys conducted by S. Tsukhlo, and 724 responded to the questioner. Most of firms, included in the sample, existed in pre-transition period. The sample is slightly biased toward machinery and chemicals in expense of fuel industry. The original sample of firms, to which the questioner was sent, is also biased toward metallurgy, but these firms had low response rate. Graph 1 compares the industrial breakdown of this dataset with the industrial composition of the Russian industry, reported in the official statistics. The GKS-total variable corresponds to the industrial breakdown, reported in the standard Goskomstat industrial statistics for 2001. The IET-total variable is the breakdown of industrial production of firms, to which the IET questioner was sent. The questioner does not have questions on output, so production data are obtained by merging the IET dataset and the Russian firms’ registry. The IET-innovations variable reports breakdown of production of firms, which responded to the innovation questioner. As in the case of the total IET sample, production was obtained by merging IET sample with the firm registry. Graph 2 reports regional breakdown of the IET sample and its comparison to the Goskomstat data. Again, both samples are more or less representative, with slight bias toward Ural region, in expense of Siberia and Far East. Such geographical bias is a natural result of the bias of the industrial breakdown toward machine building sector, because a large percentage of Russian machine building is located in Volga and Ural regions.
The IET questioner contains questions about types of innovative activities, goals of innovative activities, sources of funding, and obstacles to innovations. The survey shows that about 87% of firms are involved in innovative activities in the last three years. This number looks too high, particularly in comparison to the official statistics (see below). It seems to be consistent with other non-official survey data. Krasnochtchekova (2000) provides data from The Russian Economic Barometer (REB) survey of innovative activities of firms in 1993-96. According to REB, the percentage of firms, which were involved into either product or process innovation in these years, was fluctuating between 58-63 percent. The REB sample is similar in nature to our sample, and the hypothesis that in the early 2000s the percentage of firms, involved in innovations, increased by about 25% in comparison with 1990s sounds reasonable. Nonetheless, the statistics on the number of firms that innovate in this dataset can be biased upward. It can happen because firms, which are not involved in innovation activities, may have lower incentives to respond than the firms, which are involved in innovative activities. If we assume that all firms, which have not replied to the survey questioner, are not involved into innovative activities, then the resulting percentage of innovative firms will be equal to 41%. As we will show below, this number is still higher than the one in the Goskomstat data.
Another characteristic of innovation activities, used in this survey, is difference in innovation rate with 1980s. Enterprisers were asked whether they think that today they innovate more or less than in 1980s. About 36% of enterprises responded that their innovation rate increased since the Soviet times.
The second data set was constructed using Russian Statistical Agency “Goskomstat” publications on innovation activities of Russian firms in 2001 and 2000. These publications summarize the results of innovation surveys, which Goskomstat conducts on the annual basis. The publication does not give a lot of details on the sample, but it appears that the original database covers about 25000 Russian firms. The sample of firms is representative for the Russian industry, and closely follows the industrial and regional breakdown of the overall Russian industry (see Graphs 1 and 2, variable GKS-innovations). The publication divides all firms on those with innovations, and others, and provides summary tables for the two groups of firms. Most information is either by industry, or by region, or both. At the end of the 2001 publication there is a list of firms, which were involved into innovative activities in the last three years. This list includes brief descriptions of innovations. The publication contains a lot of other information on firms, involved in innovative activities, such as sources of finance, total spending on innovation activities, etc. This data, however, are available only in summary tables, and not available on the firm level.
The percentage of firms, involved into innovative activities in this dataset is 8.5% in 2000 and 8.7% in 2001. Since these numbers correspond to innovation activities during one year, they are not highly inconsistent with the findings of the IET survey. Assuming that each enterprise introduce innovation once every tree years, Goskomstat data suggests that about 25-30% of firms innovated in the period 1991-2001. This number is below than the number, obtained in the IET questioner. However, the IET questioner referrers to a slightly later period (the questioner was sent out in September 2003). The time period, considered in the IET dataset, is characterized by higher and more stable growth rate, then the period, considered in the Goskomstat survey. Such better economic situation can explain why IET innovation rate is higher than the Goskomstat one. We also can not exclude the hypothesis that Goskomstat’s number is biased downward because of underreporting. While in the case of the first dataset, firms, which have no innovations, had little incentives to respond to the questioner, in the case of the Goskomstat dataset, firms, which had innovations, but did not fill in the form can be mixed with the firms, which did not have innovations. This type of underreporting is very common in the Russian statistics. Goskomstat’s questioner is much larger than the IET questioner, and includes not only qualitative questions about presence of innovations, types of innovations, problems with innovations, and so on, but also rather detailed questions about the percentage of funds spent on innovation activities. Russian firms can be more reluctant to answer the questions about finance than to qualitative questions. As a result, the percentage of positive answers to the Goskomstat questioner can be smaller than for to the IET one. In addition, Goskomstat seems to first ask whether enterprises innovate, and then asks to choose proper innovation activity from the list. IET questioner go directly to the list of innovations. It is possible that due to this difference in methodology, some minor innovations are not taken care of in the Goskomstat survey, but appear in the IET survey. Graph 3 presents the ratios of the share of firms, which innovate in each industry over the average share of innovative firms in overall sample. So, this graph compares innovation rates across industries. The variables were constructed in such a way, which allowed comparison between the two datasets. Surprisingly, in the IET dataset innovations are relatively equally distributed across industries. In the Goscomstat innovation sample, firms in chemicals, machinery and metallurgy innovate more often than enterprisers from other industries.