Agricultural concerns: Limpopo has a thriving agricultural sector, but farmers told the hearings that changes to the Constitution to allow for land expropriation without compensation will lead to job losses. Picture: SUPPLIED
Agricultural concerns: Limpopo has a thriving agricultural sector, but farmers told the hearings that changes to the Constitution to allow for land expropriation without compensation will lead to job losses. Picture: SUPPLIED

This is a paper published by the Centre for Development and Enterprise


On 27 February 2018, the National Assembly made a landmark decision to review Section 25 of the Constitution of the Republic of South Africa in order to cater for the principle of land expropriation without compensation (Parliament of the Republic of South Africa1 , 2018).

This follows the decision by the African National Congress (ANC) at its December 2017 conference, where it indicated that it would start the process towards a constitutional amendment of Section 25 to make possible land redistribution without compensation, provided that it is sustainable and does not harm the agricultural sector or the economy.

Furthermore, the ANC argued that the proposed approach to land reform would be guided by sound legal and economic principles, and would contribute to the country’s overall job creation and investment objectives.

The proposal of land redistribution without compensation somewhat marks a shift in policy, and comes at a time when land reform (through both the State and market) has made more progress than experts and policymakers care to admit.

What does the data say? If we collect the official numbers of land restitution and land redistribution programmes presented by Minister Nkwinti in Parliament in 2017, the following becomes evident:

• Total area redistributed: 4 850 100 ha

• Total area restored via restitution claims: 3 389 727 ha

• The equivalent area for which financial compensation was chosen: 2 772 457 ha

Thus, a total of 11 million hectares (ha) have been redistributed through government programmes. In addition, the State has also been buying farms and, according to recent statistics, it owns more than 4 000 farms, which presumably make up the 4,027 million ha owned by the State reported in the 2015 State Land Audit. The total land area that moved out of ‘white’ ownership through the restitution and redistribution programmes as well as State procurement or State ownership (outside the communal areas) thus is effectively 15,039 million ha.

But this is not all. Earlier results from provincial land audits, as well as research by the University of Pretoria, revealed there are also many private transactions in which black individuals have bought land from willing sellers (mainly white farmers) outside of the formal government programmes and which are not included in the data presented above (Kirsten, 2018).

Fortunately, an attempt has been made to account for private land purchases by the University of Pretoria researchers (Kirsten, 2018). If we assume that, for every hectare of land redistributed through the Department of Rural Development and Land Reform (DRDLR) redistribution programme, another hectare is bought in the land market, it would suggest that private land purchases would amount to 4,8 million ha that should be added to the total land redistributed since 1994. However, a more conservative estimate where we assume that private land purchases are half of the land that DRDLR redistributes would lower this figure to 2,4 million ha that would have been transferred to black individuals through private market transactions.

This implies that a total of 17,439 million ha have been transferred from white ownership since 1994, which is equal to 21% of the 82,759 million ha of farmland in freehold in South Africa. Land reform, with the assistance of the market, has, therefore, moved us closer to the 30% target than what is commonly believed. Obviously, much more can be done to make these land reform farms commercially viable operations, however.

These numbers call into question the new policy proposals mentioned earlier. This is apart from the uncomfortable truth that the proposals contained in the various documents and bills often ignore the realities of farming in South Africa – they imply that most farms generate decent returns that can be distributed amongst many participants. In actual fact, only 4% of all farms in this country generate a turnover of more than R5 million.

In addition, most farms are in debt, and returns on equity are low – estimated at 6% in a good season. Often most of these farming operations will have large debt – anything between 30% and 50% of the asset (land) value. In 2016, farm debt, which has been rising by an average of 2% a year since 1980, had reached R145 billion, its highest level ever in both nominal and real terms, and it is estimated to have increased to R160 billion in 2018 due to the previous season’s drought (see Figure 1 below).

Figure 2 illustrates the composition of the R144 billion debt in 2016. The commercial banks held a lion’s share of 62%, followed by the Land Bank with a share of 27%, agricultural cooperatives with a share of 7%, and the rest was held by private persons and other financial institutions. The 2017 and 2018 figures will more likely present a similar picture, as the shares have not changed much over the recent past.

There is no such thing as ‘expropriation without compensation’

It is quite disheartening that the motion to review Section 25 of the Constitution disregarded the facts about the real progress that has been made with land redistribution and restitution and ignored some of the financial realities of farming. But these are not the only problems with the proposed approach, which, ironically comes at a time when Zimbabwe, one of the countries that have gone down a similar path, has established a Compensation Committee under its Land Acquisition Act to allow for dispossessed white former commercial farmers to be compensated for land seized 18 years ago. It also raises the question why the ANC is taking a position that its revolutionary counterparts from across the Limpopo are departing from. 

Nonetheless, if the Zimbabwean experience (which, though different from the ANC’s apparent approach also shares some ‘family resemblances’ with it) is not sufficient to proffer some fundamental lessons for South Africa, then it would be prudent to point out a number of facts that should compel policymakers to reconsider the December 2017 policy decision.

With the benefit of hindsight, what the Zimbabwean experience tells us is that expropriation without compensation is a catastrophically bad idea. The Zimbabweans might have seized the land without compensation 18 years ago, but they collectively paid for it through eight consecutive years of economic decline that led to job losses, de-industrialisation and a loss of agricultural export revenues. In 2009, economist Eddie Cross estimated the cost of Zimbabwe’s land reform at US$20 billion – which included lost export revenues, food aid imports and economic growth foregone, and which could have sustained Zimbabwe’s once promising economy.

Ensuing unemployment rates of over 90%, and tepid growth over the recent past, are forcing the Zimbabwean government to go back to correct the fundamental mistake it made 18 years ago – which is to compensate farmers, whose estimated compensation costs are set to amount to US$11 billion. The moral of the story is, if the government declines to compensate its commercial sector for land improvements – at the very least – then someone else will have to pay for it indirectly. The compensation effect, as we would like to call it, will see the entire economy and its citizenry paying for land seizures through lost agriculture export revenues, lost job opportunities, loss of confidence – which would lead to the flight of domestic and foreign direct investment from the sector and the economy at large.

Let us unpack the compensation effect within the South African context. There are two immediate points that are worth noting that speak to both the difficulty in implementing expropriation without compensation and the implications thereof.

Firstly, if the Constitution is amended to allow for land to be expropriated without compensation, how would the law cater for the assets on the farm and improvements made on the land? The land on its own is roughly 10% of the total value of a typical farm operation, if fixed (immovable) and moveable assets are taken into account. Would sunk investments (such as general farm infrastructure and other investment assets such as farm machinery) – which are 90% of the value of the farm – be subject to expropriation without compensation too? If compensation is due for farm assets, and not for the land itself, then the technical argument that arises is: Would it be prudent for the government to pay 90% in compensating farmers for improvements to the land in order to obtain the 10% that represents the actual land value.

Secondly, there is the complication that South African agricultural land is heavily indebted: farm debt that is linked to the actual land through title deeds that have already been used to secure loans. In this case, two scenarios are worth considering if expropriation without compensation becomes reality. One scenario is how the government handles heavily indebted land – the question here is: if compensation is not due to farmers, would there be compensation to banks, which are de facto partial owners of that land through debt? If government exonerates itself from compensating the banks, this would translate to R160 billion wiped off the books of the banks.

Another scenario is if the government commits to cover debt associated with land, which per definition becomes expropriation with compensation. The only difference is that the compensation goes to the bank that is owed money, rather than to the farmer. Let us assume that the government is sensible enough to compensate the commercial farmer for improvements made to land on the one hand, and the bank through debt owed by the farmers on the other.

If it so happens that the government determines the value of infrastructure and investments on the farms, and then uses that same value to cover the debt that is owed to the banks, then there are situations that could arise where farmers receive “zero compensation”. There might also be situations where seized farms are insolvent, in which case the government would have to pay the banks the balance of what is owed by the farmers whose land they are seizing. This scenario is already permissible under the current constitution and does not require an amendment of any law.

Thirdly, government will awaken to the realisation of the extremely complex technical headache of expropriating land without compensation, by which time land reform will have stalled altogether. This will lead to another wave of impatience that will seek to implement further draconian reforms to allow the government to seize land with impunity. We saw this in Zimbabwe, when commercial farmers took the Zimbabwean government to court over land seizures. The courts were inundated with litigation that would have taken a generation to resolve, and then, in another moment of madness in 2003, the Constitution was amended to nullify all those cases brought to the courts by commercial farmers. In that instance, the Zimbabwean government wanted to get rid of the headaches that emerged from land seizures and, in that thoughtless moment, wiped off US$10 billion in land value.

With the benefit of the Zimbabwean experience, most of which people are quick to ignore and dismiss, we learn an important lesson that needs to be the hallmark of land reform thinking in South Africa. Expropriation without compensation can cause irreparable damage to the land market by effectively reducing the value of land and sunk investments and assets. The increased risks of future expropriation without compensation means that there is likely no new capital that can come to invest further on the land. As a result, even though the costs of compensation for land fall, and thereby assisting government to expedite land reform, such costs will be borne by land reform beneficiaries who will have to be subjected to falling land prices, low onfarm asset prices and higher costs of borrowing. The latter will outweigh the former.

With government’s low compensation costs being out-weighted by the beneficiaries’ high borrowing costs and low land prices and farm assets, it becomes evident that the concept of a value-neutral expropriation is a myth. There is no such thing as expropriation without compensation in a quasicapitalist economy, because what the government refuses to pay in compensation will effectively be paid for through the negative effects that the beneficiaries experience, as well as the ripple effects in the wider economy, as discussed. The history of land expropriation under apartheid has left a deep wound in South African society, which indeed ought to be corrected. However, the enduring principle of equitable and just (not necessarily market-value) compensation in contemporary economics serves as an important reference point.

If government seizes private property, someone somewhere within the economy will have to pay, whether directly through loss in current and future on-farm job opportunities as well as export revenues, or through protracted economic decline that will erode the purchasing power of money, lead to losses in pensions and savings, and cause de-industrialisation that will destroy future economic growth and offfarm job opportunities for the current generation.

With this in mind, we are cognisant of the ANC’s view to make possible land redistribution without compensation, provided that it is sustainable and does not harm the agricultural sector or the economy. 

What could be an option to effect land reform while sustaining growth in the agricultural sector?

Having reflected on the likely unintended consequences of expropriation without compensation, it is worth relooking at some of the existing proposals that were never fully tested as means to facilitate land redistribution. These include the resolutions from the DRDLR National Reference Group (NAREG) process, the High-level Panel Report, Operation Phakisa, as well as a variety of private sector and academic proposals, amongst others.

Appreciating the fact that the current land policy proposal arose from frustration with the perceived slow progress, we have shown that there has been progress if one views this process in terms of hectares moved from white farmers to black farmers (though questions about the productivity of the use of redistributed land remain open).

With that being said, the ongoing land reform discussions provide a window of opportunity to share ideas on how we imagine the land reform process going forward. In other words, after having highlighted the unintended consequences of expropriation, one can also use this opportunity to share views on the best practice to acquire agricultural land for redistribution.

In June 2017, we argued that land reform processes should be more aligned with the ideas raised in chapter six of the National Development Plan, as we believe it has more practical steps for effective and productive land reform.

The National Development Plan suggests that the identification of transferable farms and beneficiaries should take place at a district level, facilitated by district land reform committees that were established in 2015. Under the auspices of district committees, a tripartite joint venture approach to land reform should be established. Farms for sale could be identified by the committee and a successful, local farmer would be appointed as mentor or co-investor to acquire new land together with a qualified beneficiary. The beneficiary should be selected only by the land reform committee to ensure a good working relationship between the two.

In acquiring the farm, the state would contribute 30% of land value in grant money to the beneficiary. Another 30% can be a loan from the state-owned agricultural bank in the name of the beneficiary and farmer, and the remaining 40% is a cash contribution from other farmers in that particular district. The contributing farmers would then be exempted from future land reform claims, and the farm could be operated via the farmers’ existing operations to ensure success.

A subsidised interest rate would need to be provided by the state-owned agricultural bank for the loan and backed by a state guarantee in the spirit of risk sharing. If farmers in districts worked together and get at least 30% of land in each district transferred to black farmers and ensured that it is utilised productively, then land expropriation without compensation would not be needed. Agribusinesses and commodity organisations would also have to provide post-transfer support and mentorship to new beneficiaries. This can be done only if there is a fair and transparent beneficiary selection; grants and loans are disbursed fast; title deeds are transferred and registered speedily; the government shares in the risk of redistributing land and developing new farming operations; and there is policy stability.

One of the most enduring and fundamental factors in the land reform debate is the trust deficit between the government and the private sector. Trust needs to be built in order to ensure the success and sustainability of the land programme and the agricultural sector.

In that spirit, we need more public-private partnerships, such as the Agricultural Business Chamber (Agbiz) and the Banking Association of South Africa’s (BASA) land reform model, the so-called Agbiz/BASA model, as well as the Land Bank and Afgri land reform model. These should be tested in order to create joint collaboration between government and the private sector and, in turn, to build trust.

The authors are grateful to Professor Johann Kirsten, for his invaluable comments on and contributions to some of the articles we drew from in compiling this essay.

This is a paper published by the Centre for Development and Enterprise

Please sign in or register to comment.