Andrew Kuhn

“Farmer Mac” A.K.A. Federal Agricultural Mortgage Corporation (AGM): The Freddie Mac of Farms and Ranches Has a P/E Below 9 and an ROE Above Most Banks



This is one of my “initial interest posts”. But, in this case it’s going to be more of an initial-initial interest post. I am in the early stages of learning about this company. And it’s likely to take me some time to get to the point where I can give as definitive a verdict on whether or not I’d follow up on the stock as I normally give in these posts.

First, let’s start with the obvious. As you probably guessed from the name, “Farmer Mac” is a government sponsored enterprise like Freddie Mac and Fannie Mae – except it operates in the agricultural instead of the residential market. The company has two main lines of business – again, this is just like the Freddie Mac business model except transplanted into the agricultural mortgage market – of 1) Buying agricultural (that is, farm and ranch) mortgages and 2) Guaranteeing agricultural (that is, farm and ranch) mortgages.

As a government sponsored enterprise operating in the secondary market for mortgages – the company has the two competitive advantages you’d expect. One, it has a lower cost of funds (on non-deposit money) because it issues debt that bond buyers treat as being ultimately akin to government debt. The same bond buyer might be willing to accept a 2.45% yield on a 10-Year U.S. Treasury bond and just a 3% yield on a Farmer Mac bond. This cost above the rate the U.S. government borrows at is much narrower for Farmer Mac than it is for most banks that make agricultural loans.

The other cost advantage is scale. Yes, there are banks – like Frost (CFR) – that have very low financial funding costs. But, these banks usually have to invest in hiring a lot of employees to build a lot of relationships, to provide customer service to retain customers, etc. that leads to a “total cost of funding” that is higher than what Farmer Mac has to pay. To put this in perspective, Frost’s deposits are basically the same as its earning assets (loans it makes plus bonds it buys). Frost has about $200 million in deposits per branch. This isn’t a bad number – it’s 1.5 times the deposits per branch of Wells Fargo and 2 times the deposits per branch of U.S. Bancorp (and more like 3-4 times the entire U.S. banking industry’s deposits per branch). Frost’s deposits per branch are pretty close to industry leading for a big bank. So, we can use that $200 million in deposits they have and assume a bank will almost never have more than $200 million in assets – because it doesn’t have more than $200 million in deposits – per branch. Farmer Mac has $200 million in assets per employee.

As a rule, one employee is going to cost you a lot less than one branch.

To put this in perspective, most banks spend more on rent relative to their assets than Farmer Mac spends on everything relative to its assets.

Last year, Farmer Mac spent between one-quarter (0.25%) and one third (0.33%) of one percent of its total assets on all of its non-interest expense. This is extraordinarily low relative to even the most efficient banks in the United States.

As an aside: there are a few U.S. banks – a very, very few – that charge so much in fees for non-sufficient funds, monthly charges on checking accounts, ATM fees, debit card fees, wealth management fees, etc. that they can offset so much of their non-interest expense with non-interest income that their NET non-interest expense is similar to Farmer Mac’s roughly 0.3% of assets.

You’re probably familiar with the DuPont analysis approach to breaking down the return on equity of an industrial company. You can take apart an industrial company’s ROE by breaking it down into margin (profit/sales), turns (sales/assets) and leverage (assets/equity).

A financial institution works pretty much the same way. You can break a bank’s efficiency down into 3 parts:

1)      How much does the bank pay in interest on the funds it has?

2)      How much does the bank pay in non-interest expense on the funds it has?

3)      How much equity does the bank use relative to its total funds?

There are some other factors – like charge-offs – which matter too. We can discuss those in a second. But, I’m going to spoil that discussion a bit here. I’m not so confident that we can know what Farmer Mac’s long-term charge-off rate will be. Because of the kind of loans it holds – largely first-lien mortgages on farms and ranches – Farmer Mac may have no losses in some years and then more losses in a single year than it has had in a whole decade. Although Farmer Mac was chartered several decades ago – it has a somewhat limited history relative to the length of agricultural cycles. The big determinant of losses in something like agricultural mortgages is really whether there has been a large build up in debt (especially borrowing against land values) in previous years. The last time there was a clear debt bubble in U.S. agriculture – the early 1980s – predates Farmer Mac’s existence. So, looking at Farmer Mac’s likely future charge-offs is a lot like looking at a bank that makes residential mortgages. Other than in the 2008 financial crisis – residential mortgages had very low charge-offs. But, there were then enough charge-offs in the years around 2008 to wipe out some banks (these charge-offs were sometimes more in one year than the bank had charged-off in the entire decade prior to that year). So, again, we find Farmer Mac is a lot like Freddie Mac.

So, let’s put aside likely charge-offs at Farmer Mac for a second. Instead, we’ll do that sort of DuPont analysis for banks I suggested above. Will Farmer Mac have a higher or lower return on equity than U.S. banks?


Here’s how we know that.

1.       How much does the bank pay in interest on the funds it has?

Last year, Farmer Mac’s total interest expense was less than 2.1% of its average assets. Looking at the same question from a different angle – we can see that depending on how short-term or long-term a Farmer Mac bond is, it pays anywhere from 1.2% to 3.3%. For example, a Farmer Mac bond due in about 4 years yielded about 2.9% when issued. Right now, a U.S. Treasury bond maturing at the same time, yields about 2.1%. That example is somewhat misleading – it overstates Farmer Mac’s cost of borrowing – because U.S. Treasury yields are lower now than when those Farmer Mac bonds were issued. Regardless, you can see that Farmer Mac’s cost of borrowing is less than 1% above the U.S. government’s own cost of borrowing. The company’s cost of funding from an interest expense perspective alone is not lower than many larger, successful banks. It is lower than some small, less successful banks.

2.       How much does the bank pay in non-interest expense on the funds it has?

This – rather than some interest advantage – is what makes Farmer Mac better positioned in terms of ROE than U.S. banks. Banks can pay so very little interest on certain deposits – basically none on checking accounts of households, for example – such that they could certainly match or beat Farmer Mac on that score. Where they can’t beat Farmer Mac is on the non-interest expense. Banks normally have low interest expense precisely because they offer a lot of services to their depositors. Farmer Mac doesn’t have depositors. It just issues bonds and buys mortgages (and does a few other things like guaranteeing mortgages). So, it has very low non-interest expense. This means Farmer Mac’s combined cost of funding compares well to many (though not all) banks. Let’s say Farmer Mac has a 2.1% financial expense on its total assets. And let’s also say Farmer Mac has a 0.3% non-interest expense on its total assets. The company’s “all-in” cost of funding is 2.4%. Let’s round that up to 2.5%. Okay, so Farmer Mac can fund itself at 2.5% and then buy assets that yield more than that.

3.       How much equity does the bank use relative to its total assets?

This is leverage. And it’s the biggest advantage Farmer Mac has over a bank in terms of ROE achievement. There are a few ways to measure how leveraged Farmer Mac common stock is. The method I’m going to use is definitely not the one the company’s management would use. But, I’d say the company ended last quarter leveraged about 35 to 1.

What I mean by this is that total assets at Farmer Mac are about 35 times the tangible equity attributable to the common stock. Banks – and Farmer Mac – count other equity (such as preferred stock) in their capital structure when measuring leverage. As a potential holder of the common stock – this isn’t relevant to you. What matters is the risk from leverage and the return from leverage in terms of assets/common stock equity. At Farmer Mac, that ratio is about 35 to 1.

What does this mean?

Let’s assume Farmer Mac can borrow at 2.5% a year “all-in” (this includes both the interest cost and the non-interest cost of their operation) and buy loans that yield 3.2%. This would give the company a 0.7% pre-tax return on its assets (3.2% - 2.5% = 0.7%). The company would then pay a 21% tax rate leaving it with a 0.55% return on its assets. Let’s call that 0.5%. So, if Farmer Mac could make a 0.5% after-tax return on its assets – what would you, the common stock holder, earn on your equity?

About 18%. We take 0.5% and multiply it by the 35 times leverage and get 17.5%. Again, let’s round that down. We’re left with a 17% return on equity.

How close is the situation I just described to Farmer Mac’s actual returns?

Pretty close. The official results for last year were an ROA of 0.5% and an ROE of 14.8%. Farmer Mac likes to report “core” earnings and “core” ROE. I’ll give you those numbers here – though, I have some problems with the fact they use these figures. From 2015-2018, Farmer Mac’s “core” ROE was 13-17% a year. Let’s call that an average ROE of 15%.

We also know Farmer Mac’s dividend payout ratio. Farmer Mac pays about one-third of its earnings out in dividends. It uses the other two-thirds to compound book value. This suggests that Farmer Mac’s returns – relative to book value per share, NOT stock price per share – will be a 10% growth rate and a 5% dividend yield.

To calculate your own likely return – and the form it will come in – when buying AGM, you would just divide that 10% + 5% = 15% expected return on BOOK value by the ratio of price-to-book you will be paying.

Farmer Mac ended last year with a book value of $49 per share. As I write this, the stock is trading at $71 a share. So, the price-to-book value is $71/$49 = 1.45 times book value. Again, let’s just round that up to 1.5 times. So, let’s say Farmer Mac will cost you about 1.5 times its book value.

This implies your return in the stock should be 15%/1.5 = 10% a year.

The current dividend yield is 3.66%. The rest of your 10% a year return will have to come from growth. Historically, Farmer Mac has grown its balance sheet by about 10% a year. And Farmer Mac is currently retaining enough earnings to keep growing at 10% a year while maintaining the same 35 to 1 leverage ratio I mentioned earlier.

Let’s talk about loan losses.

This is part of the complication with a bank’s reported earnings versus what it is actually earning – in cash – each year. A bank makes adjustments to its allowance for loan losses. This change in allowances may not match actual charge-off ratios, delinquencies, etc. In theory, allowances are supposed to be higher than charge-offs in good times and this will smooth out reported results. In practice, allowances tend to get raised higher after charge-offs start becoming a problem. So, allowances are less helpful in highly cyclical forms of lending than you’d think.

How high are Farmer Mac’s charge-offs in agricultural lending?


Become a Focused Compounding Premium Member.


To get the first 2,000 words of every stock write-up I do, just enter your email below.

Games Workshop (GAW): The Wide Moat, Formerly Mismanaged Company Behind “Warhammer 40,000” Has Always Had Great IP And Now Finally Has the Right Strategy