In Good Times and Bad

By Ben Duncanson
July 24, 2015

In Good Times and Bad

It has always been the mission of the Farm Credit System to make sure there is a consistent source of credit available to agriculture. With the likely down turn in the U.S. agricultural economy over the next few years, it is even more important to recognize our goal to serve farmers and ranchers in both good times and bad. We will examine this goal by looking at agricultural lending patterns over the last 40 plus years. The focus will be whether the lending patterns of agricultural lenders are pro-cyclical, counter-cyclical or constant.

Data Source

Due to limitations in the data that is available, this research is based on USDA-ERS Farm Income and Wealth Statistics data, and not FDIC and FCA call report data.  This was for two main reasons. First, this data set allows us to directly compare farm lending patterns between Farm Credit and other agricultural lenders. Second, this data set allows us to look at this trend over a longer period of time. Looking at the data over a 40 year period will help account for boom-bust cycles within the U.S. ag economy.

Methodology

The 1970 start date chosen for the research  will account for the two most recent boom cycles (1970-1974 & 2009-2013) and bust cycles (1980-1986 & 1996-2002) in the U.S. ag economy.  Also, in 1971 congress passed The Farm Credit Act of 1971 (12 U.S.C. §§ 2001–2279cc) which gave Farm Credit its modern lending authorities for eligible borrowers which, with amendments, remains the authority that we act under today. From 1971 on, this then gave Farm Credit its modern pool of eligible borrowers in the Farm Sector Debt market, which is the area that we will be testing going forward.

In order to make this easier to understand for those unfamiliar with economic analysis, I decided to limit my analysis to looking at the correlation coefficient between certain economic and debt indicators and patterns of lending from different agricultural lenders. A correlation coefficient has a value between one and negative one, with a zero value having no correlation between the variables being tested.

An easy way to think about the value of a correlation coefficient is if you were to think of two lines on a graph. If the line trends follow each other closely they will have a correlation value that would be close to positive one. If the lines move in the opposite direction, then they will have a correlation value closer to negative one. What this means is that these two variables are somehow related to each other, or are correlated. If the lines don’t follow each other at all, then they will have a value close to zero and are not correlated.  The goal of the following research is to provide a technical but simple to understand summary of some of the trends that we have found while examining this topic and only represent some of the highlights that we have found during our research.

Research

(Note: that the graph below of debt to asset and debt to equity ratios is a perfect example of two variables that are correlated and have a correlation value of .99.)

The graph above shows market share of the agricultural sector debt market that is held by different agricultural lenders from 1970 to 2013. The two largest lenders over this time period in terms of total volume were Farm Credit and commercial banks. When you calculate the correlation coefficient between the percent market share between these two lenders over the 40 plus year period you find that they have correlation value of .073, which in terms of gains and losses in market share means these two lenders are not strongly correlated.*

However, there is correlation between the lending patterns of different types of institutions and changes in the ag economy. The three indicators that were tested were: USDA’s net farm income calculation for the total U.S., debt to asset ratio, and debt to equity ratio, shown in the two following graphs.

Over the past 43 years we have seen a large decrease in both debt ratios from their height in the mid-1980s.  Also during this time period we have seen the gap between these two ratios decrease. This is most likely a result of increases in asset values, which is also reflected in the net farm income value below.

From 1970 to 2013 we have seen a large increase in net farm income in terms of nominal dollar increase (meaning dollars not adjusted for inflation), with the largest increase happening during the most recent growth period starting in 2009. This has mainly been driven by increased commodity prices over this period, with crop and livestock receipts accounting for the largest share of gross cash income.**

Note from the graph below: even though we have seen an increase in the total amount of net farm income and a decrease in debt ratios, we have also seen an increase in the total dollar value of agricultural lending over the same period. This is because farmers have taken on loans during this time for a variety of reasons such as expanding their operations and making other forms of capital investment.

Results

Now that we have talked about the variables that were tested, this brings us to our main point. With the likely down turn in the ag economy over the next few years, what are we likely to see in changes of lending patterns of the largest agricultural lenders? For this we turn to the correlation coefficient table below.

All the values in the table above are based on changes in market share of different ag lenders over the last 40 plus years. As a result, the values in the table are based on numbers provided in the first 3 graphs and not the total value of lending. This was done to better control for nominal changes over time, such as changes in the value of the dollar, changes in value of total agricultural debt market etc.

By looking at the correlation coefficient values we can see that there is a natural division in the lending pattern of different agricultural lenders. Take for example the correlation values between commercial banks and FSA compared to Net Farm Income. Commercial Banks have a value of .899 and FSA has a value of -.519, which is an example of pro-cyclical and counter-cyclical lending, respectively. This implies that with an increase in farm income, we are likely to see an increase in commercial banks market share and a decrease in FSA’s. The inverse is true when you test the correlation between ag lenders and an increase in the debt to equity ratio. Shown in the above table, FSA had a value of .776 and commercial banks had a value of -.783, these numbers indicate when there is a high level of debt to equity there is an increase in FSA’s market share and a decrease in the market share held by others, such as commercial banks.

Although these values point to cyclical lending patterns for all of the major ag lenders, there are also other economic factors that must be taken into account. This would explain a given amount of cyclical correlation value between lenders, for instance, farmers’ propensity to manage debt during different parts of the economic cycle. When farm income is higher we are more likely to see farmers pay off their debt or take on debt to expand their operations or make other capital improvements. Whereas when farm income is low or there is a large increase in their debt ratio, we are more likely to see farmers take on loans to cover their operating expenses as working capital decreases. A natural example of this is shown in FSA lending patterns, because FSA loan programs are designed to help farmers when other forms of credit are not available to them. This results in FSA acting as a natural countercyclical lender driven by its design in federal farm policy. The same is partially true for individuals and others lenders. However, because this variable is made up of many different types of lenders, it is hard to determine an overall pattern in lending.

When looking at the correlation values above for Farm Credit, we find that Farm Credit has a consistently low correlation value. What we can infer from this is that Farm Credit’s lending patterns are less correlated with cyclical changes in the ag economy than other lenders, and therefore Farm Credit acts more as a primary and constant source of credit to the farm debt market.***

This is illustrated best by looking at the last downturn in the ag economy that happened from 1996 to 2002. The table below shows the correlation values for the decrease in net farm income for the same four groups of agricultural lenders as the first table but is limited to the 1996 to 2002 period. During this period Farm Credit had a correlation value -.401, which means that even though there was a decrease in total U.S. net farm income (from $58 billion in 1996 to $39 billion by year end 2002) Farm Credit continued to be one of the leading lenders to farmers, ranchers and other agricultural producers and actually increased market share as farmers turned to Farm Credit as a reliable source of credit.

Conclusion

Given that it is Farm Credit’s mission to provide a reliable source of credit to farmers and ranchers during good times and bad, it is not suspiring to find that Farm Credit’s lending patterns are not strongly linked to cyclical changes in the ag economy. Over the past 40 years Farm Credit has helped many farmers, ranchers and other agricultural producers operate and grow their businesses. As we move forward, Farm Credit will continue to perform on its mission to support the future of American agriculture in both good times and bad.



* This also means that in terms of gains and losses in market share between Farm Credit and commercial banks, these two institutions’ lending patterns are independent of each other, so every new percent increase in market share that the Farm Credit System gains does not mean that the commercial banks then lose that percent of their market share. This is due to the high level of competition between lenders for every new dollar of agricultural debt during an expansionary period in the debt market.

**Note: net farm income is calculated by USDA Economic Research Service (ERS) by adding the total number of crop receipts and livestock receipts to the total value of government payments and other forms of cash income together to get the total gross cash income. Total dollars of gross cash expense, including interest expense, are subtracted from gross cash income to determine net cash income. Net Farm Income is calculated through the same process, but an accrual adjustment is made to account for net changes in inventory to account for year to year carry over and other changes in non-money income for the following year. Then depreciation charges and allocations for hired labor plus other expenses are added to gross cash expense to yield total farm expense, these adjusted values are then added to income and expense respectively in order to calculate net farm income. Net farm income was chosen for this analysis because it more acutely accounts for incremental changes in farmers’ year to year balance sheets and because it takes into account a greater number of factors that affect the ag economy, such as changes in labor and other input values.

*** Note: when testing these variables against total volume of ag lending, Farm Credit also had the correlation value closest to zero, as well as when tested against other farm income, debt and economic  indicators.

 


About Gary Matteson
Gary Matteson knows agriculture first hand. Until recently he was a small farmer operating a greenhouse business in Epsom, New Hampshire. Matteson now works at the Farm Credit Council, the trade association for the nationwide Farm Credit System. He is an advocate for young, beginning, small, and minority farmer outreach programs. Matteson is responsible for spreading best practices for beginning farmer lending and training among Farm Credit Associations, generating new program ideas to benefit them. In addition to working directly with farm groups, Matteson is active in policy related to new entrants to farming. He now serves on the USDA Advisory Committee on Beginning Farmers and Ranchers.