esti-66
mates are used to analyze sensitivity and to generate confidence intervals for each of the decom-posed measures.
67
The mean debt capacity utilization was roughly the same level across the sectors in 1996, with a mean utilization of 0.62 for crop, 0.58 for dairy and 0.66 for pigs. The mean debt utilization and the results from the decomposition of the DDi are presented in Table 2.3 for the years 2004 to 2009.
The DDi and the components are listed on the right-hand side.
Table 2.3.
Mean debt capacity utilization and Debt Development index decomposition from 2004 to 2009 Year
Mean debt capacity utilization
No. of farms
both years*
Debt Devel-opment
index
Change in debt capacity
utilization
Scale related change in debt
capacity utilization
Change in debt capacity
Scale related change in debt capacity Crop
2004 0.61 64 1.346 1.081 0.938 1.296 1.048
2005 0.62 64 1.617 1.113 0.931 1.500 1.065
2006 0.52 57 1.436 0.911 0.939 1.608 1.046
2007 0.46 55 1.515 0.817 0.920 1.923 1.093
2008 0.50 46 1.654 0.918 0.919 1.875 1.070
2009 0.50 35 1.958 0.855 0.965 2.334 1.008
Dairy
2004 0.61 617 1.496 1.114 0.919 1.365 1.134
2005 0.60 538 1.532 1.098 0.919 1.385 1.154
2006 0.57 437 1.663 1.061 0.899 1.586 1.163
2007 0.54 364 1.772 1.021 0.918 1.756 1.181
2008 0.56 295 1.929 1.011 0.905 1.880 1.198
2009 0.63 152 2.385 1.176 0.905 2.137 1.139
Pigs
2004 0.78 265 1.562 1.252 0.937 1.316 1.070
2005 0.76 231 1.675 1.142 0.915 1.503 1.122
2006 0.74 252 1.757 1.133 0.911 1.624 1.112
2007 0.76 143 2.157 1.168 0.909 1.854 1.149
2008 0.80 99 2.501 1.221 0.941 2.051 1.090
2009 0.79 87 2.645 1.204 0.941 2.391 1.151
* with 1996 as the base year for farmers in the dataset in 1996 and the year in question.
The change in debt capacity utilization is below 1 for crop, close to 1 for dairy and around 1.2 for pigs. This reflects the development in the mean debt capacity utilization which fell to 0.50 for crop, increased to 0.63 for dairy and increased to 0.79 for pigs in 2009. The debt capacity utilization can be interpreted as the change in debt to debt capacity ratio, which is a good measure of credit re-serves and is closely related to the debt to asset ratio when controlling for management, age, and the
68
lender’s valuation of assets. The scale related changes for both debt capacity and debt capacity utili-zation range from 0.899 to 1.198 for all three subsectors, which suggests that the change in scale does not cause changes in debt structure. The main measure in Table 2.3 is the debt capacity change, which roughly doubled in the period, also shown in Figure 2.3. This reflects an outward move of the VRS debt possibility frontier.
The interpretation of the numbers in Table 2.3 can be illustrated by the following example: Suppose you were a pig farmer in 1996 with total liabilities of €1,000,000 and an estimated maximum debt capacity of €2,000,000 based on the most indebted comparable farms. Your debt capacity utilization was 50% and your credit reserves were estimated at €1,000,000. In 2009, the change in debt capaci-ty compared to 1996 was 2.391, which means that the maximum debt capacicapaci-ty increased to 2.391 x
€2,000,000 = €4,782,000, holding assets and management level constant, adjusted for inflation. If the debt of your farm remained constant at €1,000,000 (adjusted for inflation), your debt capacity utilization would be 1,000,000 / 4,782,000 = 20.9 % and your credit reserves would be €3,782,000.
Now suppose that your debt capacity utilization followed the sector level change in debt capacity utilization, i.e. 1.204. Then your debt capacity utilization in 2009 would be 50% x 1.204 = 60.2%.
This means that the total liabilities would be €4,782,000 x 60.2 % = €2,878,764 and your credit reserves would be €1.903.236.
The results in Table 2.3 are the means of the scores for farmers in the sample in 1996 and the year in question. The measure is biased because it is only calculated for the farmers who were farmers in 1996. Operator age is used as an input and the farmers in the measure for 2009 are older than the average farmer. Also, some farmers start with a small farm and build up a larger farm. Hence it is expected that the debt capacity change is measured at the part of the frontier where the farmers are older and where the farms are larger. Finally, the number of accounts included in the mean calcula-tion decreases with time.
The debt possibility set in 2009 consists of 683 farms (see Table 2.1) for the crop producers, and the efficiency estimates are calculated on the basis of this debt possibility set. The average DDi change for crop producers is estimated based on the 35 farms in the data set in both 1996 and 2009. These farms are not representative, as the farmers had been in farming for at least 14 years in 2009, which precludes young farmers. This may constitute a survivorship bias. It is important to note, however,
69
that this is not a bias in the estimation of the frontier. As all data are included in this process, the possible bias is because development is not measured homogeneously on the frontier. A solution to this issue is to measure the mean farm. Each input and output in the mean farm is the mean of all farmers for the relevant year within the production type. This average farmer does not age. The re-sults for the mean farm show no significant difference, which suggests that the bias is minor. Meas-uring the development at ‘interesting’ places on the frontier may be seen as tantamount to the prac-tice used in Logit and Probit analysis where we want to measure response to marginal changes in at ‘interesting’ values of , which is often, but not always the mean (Wooldridge, 2002).
The change in debt capacity generally drives the change in the DDi over the years. Scale effects tend to be minor, while the debt capacity utilization has some positive effects on DDi for dairy and pig farms and a negative effect on DDi for crop farms. The ability to raise debt increased from 1996 to more recent years, which is displayed in Table 2.3. This ability was primarily utilized by the Danish dairy and pig farmers, while crop farmers seem to have been self-imposing more restraints on the use of credit as change in debt capacity utilization was above one for dairy and pig farmers, but below one for crop farmers in 2006 to 2009.
2.5.2 BOOTSTRAP RESULTS
Bootstrapping makes statistical inference possible based on the empirical distribution of the boot-strap estimates of the decomposed index. The bootboot-strap is used to identify the number of farms for which there is a statistically significant debt capacity change at the 5 per cent level. Figure 2.4 shows the share of farms for which there is an expansion in the debt capacity at the 95 per cent con-fidence level.
In summary, the results show that ease of access to credit increased significantly from 1996 up to the global financial crisis with the magnitude of the increase being roughly double. Our debt capaci-ty measure shows increases in the magnitude of 76 - 92 per cent from 1996 to 2007. From 2004, all three subsectors have a large proportion of farms with a significant increase in the debt capacity (change in debt capacity > 1, at the 95% confidence level). There is no doubt that access to credit was easier in Danish agriculture in 2007 than it was in 1996. What is puzzling about our results is that access to credit seemed to increase during the financial crisis in 2008 and 2009. Our measure does not distinguish between debt generated for investment finance and debt generated to cover
70
operating losses. The development in debt is especially difficult to track in transition periods early in a crisis when investment finance becomes constrained, but current operations are financed (as lenders consider this to be the loss minimizing strategy). With some differences across the subsec-tors, many farmers incurred losses on financial arrangements as well as losses on current operations during the crisis. These farmers experienced increasing debt capacity (utilization), drawing on their credit reserves, even though they were experiencing decreasing access to investment finance and were possibly credit constrained in executing profitable investments. The measure does not distin-guish between the behavioral or strategic reasoning behind loan approvals or credit expansion.
However, we believe that Danish farmers experience greater credit constraints now, than they did prior to the financial crisis.
Figure 2.4. Share of farms with a significant increase in debt capacity since 1996
* Due to computational burden, results for dairy are not reported for 1997.