Every so often I get asked about whether it is valid to include improved land values in economic evaluations of land conservation works on farms, or I read an analysis in which it has been done. Both of these happened recently, so it is probably time to draw attention to my notes on this issue again.

The way it is usually done, or proposed to be done by non-economists, amounts to counting the same benefits twice. Your starting point should always be not to do it.

It is possible in some special situations to include some consequences of changes in land value as a benefit in an analysis of a land conservation investment, but you need to be very careful how you do it in order to avoid double counting. For the purposes of this sort of analysis, land value differences would reflect differences in future productivity as a result of the investment in conservation. If you do include changes in land value at a particular date, then you must exclude from the analysis all productivity changes subsequent to that date, because they have already been included in the land value change. (Rationale: to realise the capital gain, you have to sell the land, and then you don’t have access to the land to realise subsequent productivity differences.)

Conceivably, you could include productivity differences for 20 years, and then a difference in land value at year 21 to factor in ongoing effects beyond year 20, provided that you then discounted that difference back to the present. In practice, however, I would ask, how should you determine the change in land value to be included at year 21? The best way would be to do a discounted cash flow analysis of subsequent productivity differences. If so, why not just extend the time frame of the productivity analysis? It amounts to the same thing.

There are some complexities that could be added to this relatively simple story. For example, if land value increases, farmers may be able to access greater debt finance to invest in productivity. Benefits from such investments, if you could anticipate them and quantify them, would be legitimate inclusions in the analysis of the investment in land conservation. However, they are really second-order benefits – almost certainly much smaller than the error margins surrounding the first order benefits – and they could only ever be highly speculative. They are also completely different in nature to the direct inclusion of changes in land value.

Be aware that if you assume that prices, yields and costs are constant over time (which is a common, though dubious, thing to do), there is no justification for even expecting land prices to increase, let alone including these increases in the analysis. Unless …

Conceivably there could be an expectation of speculative increases in land value, not reflecting productive value, and a belief that such increases would occur WITH the land conservation investment but not WITHOUT it. If one has some sound rationale for expecting this, then fine, include it as a benefit, but I wouldn’t like to be investing any of my own money on that basis, and good luck trying to justify it to any reviewers of your analysis. Substantial differences between the value of land for production and the amount people actually pay for it are not sustainable in the long term, due to competitive pressures, unless the land has values other than for production (e.g., it is close to a city).

Putting that aside then, I would note that the analyst should be considering whether the assumption of constant prices, yields and costs is the most reasonable. We know that real prices for agricultural commodities have been falling for at least the past century (the real price of wheat has been falling since the 1700s), and I don’t think it’s time to expect that trend to change. We also know that yields have increased, and will probably continue to. The “constant real everything” assumption is only reasonable if price falls and yield increases are expected to cancel out.

Land prices increases COULD occur if technology is advancing fast enough for productivity growth to more than offset the price falls. This has been happening in Western Australia, though it applies to land generally, not just land that has had a land conservation investment.

If the productivity growth is different in absolute terms with and without the land conservation investment, then you perhaps need to worry about it in the analysis. For example, if degraded and non-degraded land both have the same percentage productivity growth – say 1% per year – this implies a small difference in absolute terms, because 1% of a larger yield is more than 1% of a small yield. You might want to accounted for this, although it smacks of desperation!

If you did include this effect of different absolute productivity improvements, I would strongly suggest doing so in the discounted cash flow analysis of the land conservation investment, rather than as a land value change. Pulling a number like 1% out of the air obscures the implicit assumptions about rates of productivity improvement etc. that are being made. Better to represent these assumptions directly and explicitly. Even if you did want to base the analysis on land values rather than productivity, it certainly wouldn’t be correct to include the full 1% increase as a benefit in the analysis of the land conservation investment, unless you believed there would be no increase at all in the value of land which had not benefited from the land conservation investment.

As always, when you are discounting, be careful in the use of real (inflation excluded) versus nominal (inflation included) rates of change. If the 1% increase in land value is in nominal terns, rather than in real terms, then this would actually be a real price fall (assuming inflation exceeds 1%), presumably implying that yield increases are not fully offsetting real price falls. But this doesn’t change the logic presented above. Including changed land value in the analysis would still depend on whether the absolute fall is different when you do or don’t undertake the land conservation investment. And again, I still wouldn’t represent this by a land value change, but by a set of differences in annual profit depending on changes in yields and prices over time.

Overall, it’s an area where it is easy to go astray unless you know what you are doing. The cases where you can validly include land values are really special cases and would not likely make a big difference anyway, so the safest approach is just to exclude land values completely and make the period of the analysis sufficiently long to capture productivity effects over the long term.

David Pannell, The University of Western Australia

Further reading

Pannell, D.J. (2004). Avoiding simplistic assumptions in discounting cash flows for private decisions, In: D. Pannell and S. Schilizzi (eds.), Discounting and Discount Rates in Theory and Practice, Edward Elgar, (forthcoming). full paper (45K)