Union Investment achieves a quantum leap with Aqua transaction
A portfolio transaction opens the window to the future. The due diligence for the Aqua-portfolio was supported by an intelligent, self-learning software.
In several industries, the processing of so-called ‘Big Data’ escalates at a rapid pace. But also in the real estate industry a lot is happening in this field by now. We have already become accustomed to the fact that real estate portals use the large quantities of data, that accumulate in the brokerage business, for prompter and better market analyses.
But now, the portfolio managers of Union Investment Real Estate together with several partners achieved a quantum leap in the transaction business.
It all started with a routine decision: Every year, Union Investment Real Estate implements a sales program with a volume of around one billion euros. The company invests the proceeds in other properties. And in 2015, it was supposed to work similarly. The selection of potential properties had already begun in mid-2014.
Since the financial crisis, the properties were sold individually. However, this time Philip La Pierre, Head of Investment Management Europe, and his colleagues decided on a different sales channel. A total of 17 office buildings from four different funds of Union, located in six different European countries, were to be sold as a package. One reason was that they were suited well for this purpose.
“The homogeneity of a portfolio is at the top of the priority list of potential buyers”, La Pierre said.
And that applies to the selected properties with respect to the cash flow and the lease durations: the WALT is about seven years, the quality of the buildings has a comparable level. The gross rental proceeds from the portfolio amount to 60 million euros.
However, the main motive was clearly the market situation. Another big portfolio of a different fund provider was just available and received the attention of many bidders.
“Part of our plan was to pick up those bidders that did not get the chance to buy there”, said La Pierre.
Then, the real work began – the data preparation for the buyer of a package. The data processing and the provision in data rooms are essential for the success of such transactions. Lack of data can quickly cause price reductions due to the associated uncertainties. It takes a lot of time and effort to get a complete overview of approximately 200 commercial lease files from six different jurisdictions. Typically, as a next step, several well-paid lawyers have to scan the contracts page by page looking for relevant clauses.
But this time everything was different. Union Investment Real Estate decided to use a ‘disruptive technology’ and commissioned the Berlin-based IT company LEVERTON to conduct the preliminary due diligence of the contracts. LEVERTON’s core competency is the development of a self-learning software that recognizes linguistic meanings and its variants, and its application to the real estate industry.
“The software belongs to the category of ‘deep learning’, which means it learns by using algorithms that imitate neural brain structures. That is why letters and contents of scanned documents can be recognized automatically”, explains Dr. Emilio Matthaei, CEO of LEVERTON.
For example, lease data, option rights, change-of-control clauses and the like. During the semantic detection of the lease files, usually present in PDF format, the software not only recognizes sentences using the same wording, but also phrases that are formulated in a different way, but in meaning have similar content – it ‘learns’.
In short, LEVERTONs software read around 200 lease files automatically, added hyperlinks to relevant paragraphs and thus made them detectable in an easy and clearly structured way.
Matthaei: “We transform Big Data into Smart Data.”
The data can be transferred into the desired ERP-system automatically. The seller has a completely new sensitivity of his portfolio.
“We knew about 99.9 percent of our portfolio in detail and went into the negotiations full of confidence”, says La Pierre.
Also the potential buyers – a total of nine parties were interested in the offer of Union Investment Real Estate – were able to enjoy a new speed of due diligence. The time of manually checking files and folders page by page is over with this new technology. The LEVERTON-software masters 20 languages so far, with more to come.
A recent study by the World Economic Forum expects that automated and digitized processes could cost around five million jobs in the industrialized countries by 2020 – particularly for simple office work with repetitive character. The LEVERTON-technique, however, does not make humans redundant, but it accelerates the legal due diligence enormously. Even the error rate that occurs during the extraction of the relevant attributes – according to Matthaei it lies between 5 to 15 percent – does not change this.
Therefore, there still need to be competent people who validate the automated due diligence. Those are usually lawyers. Thus, it would not be surprising if law firms would react with skepticism and resistance against the transformation of stacks of files into Smart Data. The law firm, Clifford Chance, that was responsible for the transaction preparation of the Aqua portfolio did nothing like that, but the contrary:
“We have been working with LEVERTON for almost two years and we like to offer its technology in addition to our services”, reports Cornelia Thaler, Partner at Clifford Chance and consultant for the Union Investment Real Estate during the Aqua transaction.
The time saving is actually enormous, she says:
“Manually a lawyer needs three to five hours to examine one lease file. If the data extraction of the LEVERTON software is used beforehand, only 20 to 45 minutes are needed.”
4.3 Percent Return
As a final result, the high level of transparency and the extensive assessment of all relevant facts in the data room have paid off for Union Investment Real Estate. The French insurance company Amundi acquired the Aqua portfolio for 1.4 billion euros.
“That is higher than the value calculated by experts”, says La Pierre pleased.
The application possibilities of self-learning software that translates documents into Smart Data are by no means limited to the transaction area. Also the facility management can benefit. For example, Strabag PFS has launched a pilot project with LEVERTON. Thereby, the computational linguistic software extracts the contractual breakdown of maintenance obligations in approximately 1,000 tenancies.
“Data is then found quickly and analyses, decision-making templates, risk assessments and reports for all stakeholders are available ‘at the press of a button’ ”, hopes Marko Bohm, CEO of Strabag PFS.
The journey to the Real Estate World 4.0 has only just begun.
The full and original version of this article was written in German by Christof Hardebusch and published in immobilienmanager, Jan/Feb 2016, pages 52-53. The online version can be found here.
LEVERTON develops and applies disruptive deep learning / machine learning technology to extract and manage data from corporate documents in more than 20 languages.Our clients from the financial and real estate sectors optimize their contract management significantly, manage operational teams more efficiently and process transactions faster. Asset documentation with an underlying face value of more than €40 billion is managed with LEVERTON’s smart data technology. The integrated smart data platform runs on LEVERTON servers in ISO 27001 and ISO 9001 certified data centers that guarantee highest security.