This Commentary reviews historic farmland returns and shows that, contrary to its reputation as a low return industry, farmland returns have out-performed equities in three out of the past five decades.

After making this “bull case” for farmland, we then argue these apparent “excess returns” of farmland are a consequence of the volatility of farm operational returns

The Commentary concludes by contrasting the (logically higher) valuations that financial investors might place on farmland with the valuations that (more constrained) family farm purchasers are likely to be able to pay.

Introduction – the danger of being a farmer at MBA school (or at a dinner party)

MBA schools characterise farming as a commodity industry, subject to negative weather events and low returns. Farmers are often told at dinner parties that farming:

  • is a low return asset class
  • which produces commodity products, which cannot easily be differentiated
  • making it a puzzle why land is so expensive
  • with the prices being paid for farmland recently being particularly crazy!
A case study of one of these “expensive” assets: South Island New Zealand sheep and beef land

Over the years these views made sense to me also. Twenty years ago $3,000 per hectare for land in our valley seemed high. By the time my father died in 2004 they had risen to $6,000. Now they have again doubled to $12,000. At each stage of this process the farm was generating only 4% operating returns on assets*.

*This must have been true in the 1970s also as I can remember my father telling me then that “NZ sheep and beef farms will always yield only 3% to 5%”. With the benefit of hindsight I believe he was suggesting that as soon as they start to yield more farm prices will rise to bring the yield back down to their equilibrium range.

This commentary attempts to unpick the puzzle of an asset class that often looks very ordinary (or in a bad year downright unattractive) when looking at the current year’s financial accounts, yet so often looks good when looking in the rear view mirror.

Historic US farmland returns

The graphs below summarise five separate decades of data on the total returns** to US farmers and also looks at returns in a couple of other countries.

**these are returns on assets, including both operational returns and capital gains.

The 1960s

In the 1960s US farmland generated 11.8 % annual growth, relative to 7.7 % for stocks and 2.5 % for US Treasuries:

Compounded Return of Selected US Assets Classes 1960-1969


Source: Map of Agriculture analysis

The 1970s

In the 1970s US farmland had a remarkable decade with annual total returns of 18.2% per annum in a period of stagflation when both equities and bonds had negative real returns. Farmland, gold, and other real assets benefited from the “flight to quality” in the face of high inflation and low growth.

Compounded Return of Selected US Assets Classes 1970-1979


Source: Map of Agriculture analysis


The 1980s

The 1980s is the one decade in our sample when US farmland under-performed both US Treasury bonds and equities.  In this decade US farmland, with 7% annualised returns, “normalised” – giving back some of the relative returns from the 1970s (when farmland valuations like gold, may have become over-extended).

Compounded Return of Selected US Assets Classes 1980-1989

Source: Map of Agriculture analysis


The 1990s

The 1990s was an infamous period for agricultural over-production caused, primarily, by excessive subsidies (remember butter mountains and wine lakes). However US farmers still did ok with an annual growth of 12% vs. equities at 17% and US Treasuries at 7%.

Compounded Return of Selected US Assets Classes 1990-1999


Source: Map of Agriculture analysis

The 2000s

Finally, in the decade to 2010 US farmland generated a compound return of 13% per annum vs equities -1%.

Compounded Return of Selected US Assets Classes 2000-2009

Source: Map of Agriculture analysis

Other Countries

A 2008 study by Eves and Painter (table below) for the period 1990 to 2005 suggests that in New Zealand (at 14% per annum) and Australia (at 10%) farmland total returns were slightly ahead of those in the US, with Canadian returns not far behind at 6% (Canada, by the way, has since picked up nicely).

Income, Capital Gain and Total Farmland Investment Yields (1990-2005)

Source: Eves and Painter 2008

The above data was gathered by the US Department of Agriculture and other national agriculture departments. The data is from operational farmers and includes both trading and capital returns to farmers (who for the most part own the land they farm).

Earnings growth as the key driver of farmland returns

In my view, and this is a view I think many farmers would share, the key to creating value in a farming investment is growing the production of our farms over time. Farmers do this within existing crop types (old fashioned “productivity gains”) and we also create gains by undertaking land-use change (“development”).

Considering the same-crop type; farmers routinely increase their annual production per hectare via technological, management and genetic gains (some costless, others involving investment).  During the green revolution (most of which occurred in the 1950s and ’60s in developed countries; and in the ’60s and ’70s in less developed countries) these productivity gains ran along by as much as 5% per year.  More recently same crop gains have fallen to 1% to 2% per year in most farming systems.

Annual productivity increases and development gains (net of expected deflation of food commodity prices, and adding expected inflation) are typically capitalised into farmland values.  Since these gains are expected to continue into the future a simple equilibrium “valuation model” should add these sources of gain to the % operating return. E.g. if a NZ dairy farmer budgets 6.5% operating (i.e. cash-flow) returns, 2.5% gains from productivity and development (net of investment) and 2.5% inflation gains then he may reasonably expect 11.5% total returns on assets.

Farmers will typically then (moderately) leverage these assets. E.g. our typical NZ dairy farmer might  budget a gross 15% total return to equity after leverage. After corporate overheads and taxes net NZ dairy total returns might then average 12% for the more highly leveraged operators, and 10% for the more conservative.

10% or 12% p.a. is quite an attractive return for a defensive, low depreciation real asset with a negative correlation with equities*** so the question arises why does the market not further “bid up” the price of farmland, to bring expected returns in line with other assets?

*** Professor Bruce Sherrick of the University of Illinois estimates a low/negative correlation between farmland and equities

Correlation of asset returns by class with Illinois farmland returns


Source: Sherrick, Mallory and Hopper 2013 *Iowa is 100%

But what about the annual variability of farmland returns?

Because of commodity price volatility and weather variability returns of most farming systems vary (sometimes significantly) from year to year. E.g. the graph below for US Kansas wheat returns shows operating returns varying from 1.4% to 7.3%.

EBIT/Total Assets for Kansas Wheat Farms


Source: Map of Agriculture analysis

Similarly NZ Waikato dairy returns show a range from 3.6 to 8.3%:

Operational Return/Total Assets for Waikato Dairy Farms

Source: Map of Agriculture analysis

In a further example operating returns of dairy farms in Victoria, Australia ranged from -0.5% to +5%:

EBIT/Total Assets for Victoria Dairy Farms

Source: Map of Agriculture analysis

Operational returns volatility as an explanation for the high historical returns of farmland as an asset class

I have come to the conclusion, in trying to unlock the “puzzle” of farmland equilibrium valuations, that the above often ferocious annual operational volatility is the reason historic farmland prices were not bid up higher (i.e. the reason returns are not lower).

Despite the lectures we get at dinner parties farmers are quite aware that “the next decade” will normally see total returns in line with or better than those of the stock market (and probably less volatile on a total return basis).  It is precisely this passionate belief in our “unattractive” industry that our urban dinner party cousins find so perplexing. However, farmers simply do not have the cash to bid up the price of land further than the normal yields at which that class of farmland trades. This is because farmers and their banks tend to finance most farmland businesses primarily with equity. For many crop types too much debt funding can be precarious with the above large annual swings in operating cash flow.  Further, intergenerational equity distributions (buying a house in town for the parents, buying out non-farming siblings) mean that family farmers are perennially undercapitalised.  The scarcity of farmers’ access to equity means projects need to generate higher returns than they would if they were able to use more debt.

Implications of this theory for financial investors

Family farmers (i.e. owner-operators and/or operators leasing land from family members) are the overwhelming owners of farmland in most countries, owning well over 90% in most jurisdictions.

A general analysis of what may happen to farmland equilibrium valuation ratios if farmers are replaced by e.g. institutional investors as the main buyers of farmland is beyond the scope of this Commentary except to say:

  1.  The above “operational volatility”, and “equity scarcity” arguments suggest equilibrium “fair” pricing of farmland should be lower for farmers (who cannot easily diversify downward spikes in returns) than for more diversified investors. In other words (and other things being equal – a big assumption hence this only a partial analysis) institutions, possessing sufficient information to model returns and risks, would bid land prices to a new equilibrium that family farmers would not be able to afford.
  2. It follows that farm operational returns (“cap. rates”) may move to a lower equilibrium (land prices may rise) when more institutional capital is entering the market.  Conversely operational returns may rebound to a higher equilibrium (land prices may fall) in periods when institutional investors cease being net buyers.
What happened in the past when financial investors became big buyers of farmland?

Anecdotally, we gather there was widespread institutional buying of farmland in the US and the UK in the late 1970s. If supported by actual data this could help to explain the expanding valuations shown in our 1970s US farmland returns history above.

  • Looking backward, farmland returns in major agricultural countries appear to have been at least as good as equities over long periods of time (and may have shown less volatility and negative  correlations).
  • A simple forward-looking model of farmland total returns sums expected earnings growth (from productivity, development and net inflation) with budgeted cash flows. At least one category of contempory farmland appeared attractive using this model.
  • However the volatility of farmland operational returns constrains the valuations that family farmers can afford to pay for land. This may help to explain it’s quite satisfactory levels of return over time.
  • Financial investors’ portfolios are generally non-correlated with agricultural risks. This suggests these investors may be in a better position to diversify farmland risks than family farmers. They are also less constrained by scarcity of capital than family farmers.
  • In periods and regions when there are high levels of “financial investor” interest in farmland, purchases by financial investors may bid up equilibrium valuations.
Further research

A more thorough analysis of the above theory would study the interplay between financial investment flows and farmland valuations over time and across jurisdictions. This research will require data on buying profiles of net farmland purchasers over time. Watch this space! Any questions or comments gratefully received.

Warm regards,

Forbes Elworthy

Published: 25 September 2014