It has been detected that many fraudsters and money launderers act on two levels: first, they set up companies that in turn set up others. In this way, they form a network of enterprises that are interrelated through their capital. On a second level, they set up companies in the name of their spouses, children, parents-in-law, children of former spouses, etc., that form another network of enterprises, this time related through their owners. This generates a chain of hidden relationships that can only be uncovered using sophisticated algorithms based on artificial intelligence.

Algorithms are extremely useful in combating tax evasion, for two reasons: they help to identify a person’s family network (information which it is not mandatory to declare) and they make it possible to aggregate the wealth an individual controls through his or her network of companies and that which the same individual controls through his or her relations.

In this context, in order to conduct the research on which this article is based, first the network of family relationships in Spain was obtained. On the basis of available data from the Tax Agency register, relationships between spouses and between parents and children were used to connect several generations. In this way, with the aid of algorithms, the degrees of kinship were reconstructed up to the third degree of consanguinity and affinity. Thanks to this, in 2017 the 87.6 million relationships that were declared for purposes of both income tax (among spouses and their children) and inheritance tax grew to almost 258 million detected relationships. This also afforded a better understanding of the links that exist between individual and family wealth.

The largest number of relationships to be found and then used to form companies was between siblings (56.78 million), followed by relationships between uncles and aunts (3.67 million), nephews and nieces through bloodline or marriage (3.6 million), brothers- and sisters-in-law (3.3 million), ascendants or descendants of the spouse (2.6 million) and cousins (1.7 million). In addition, there are 15 types of undeclared relationships with more than 500,000 cases, the last of which, between former brothers- and sisters-in-law, accounted for 505,575 cases.

Parallel to the above, it was also possible to map the network formed by relationships of investment among Spanish firms by using data from corporate tax returns and the register of entrepreneurs and professionals. In this way, it has been revealed for the first time that almost 2 million taxpayers are involved directly as shareholders or owners in almost 3 million corporate entities, in the form of both public limited companies and private limited companies. Their investment is direct in 1.3 million of these firms, whereas in 1.7 million cases the investment is through other companies.

It has also been observed that some individuals are shareholders in more than one enterprise, and that the person with the most shareholdings is involved in 382 companies. On analysing this network and measuring the distances between individuals and companies, it has been revealed, for example, that 14,812 people own part of a commercial enterprise through the procedure of holding shares in one company that invests in a second company, which in turn invests in a third one, which again invests in a fourth one, in so-called “investee companies”. Furthermore, the algorithms detected that some entrepreneurs have shares in as many as 15 investee companies.

All this information served to reconstruct the network that connects individuals to their wealth, by adding to what they possess individually what they also own in corporate holdings, regardless of the number of investee companies. These companies may be set up for the purpose of obtaining a tax advantage or simply to conceal assets.

The third step of the research was to calculate the net assets of each company and each individual. It also succeeded in exposing who controls the companies, a key element for remedying the fraud or crime committed.

In 2018 the Spanish Tax Agency register contained 71.06 million taxpayers, of which 62.67 million were natural persons and 7 million were legal persons. There are more natural persons than inhabitants in Spain (46.5 million in 2017), due to the fact that the register also includes non-residents (more than 9 million) and deceased taxpayers still within the statute of limitation period.

In 2017, of the 46.53 million inhabitants registered in Spain, 202,437 (0.43% of the total) declared assets to the value of 669,062 million euros in their wealth tax returns. However, this figure does not represent the totality of their wealth. These taxpayers have more than 2 million euros and are therefore obliged to file this type of tax return.

Using the rest of the information available to the Tax Agency, and thanks to algorithms and big data techniques, a more exact figure was obtained. In this way, it was detected that some 37 million citizens own an asset of some sort, such as the bank account where they receive their salary. This represents 79.5% of the population, with an average positive net worth of 141,000 euros per taxpayer. It was also found that 15.3 million registered natural persons, of whom 9.23 million were living Spanish citizens, possessed no assets of any sort in 2017.

For the first time, the calculation of wealth distribution has included wealth held in companies not listed on any stock exchange, whether the taxpayer has a direct shareholding in those companies or he or she is involved in them indirectly through other firms. This represents 11.37% of total wealth in Spain. This information has been unavailable to date, and without it the estimated wealth distribution was inaccurate, as it failed to appreciate the real economic capacity of the richest sector of the population. Knowing this information means, in practice, being able to observe the distribution of total net worth. In this respect, it is also important to note that the larger the personal fortune, the more it is accumulated in firms that are not listed on the stock market, a mechanism that it is difficult for the Tax Agency to detect, and as such, conducive to tax evasion and fraud.

Once the network of relationships between individuals and companies, and between companies, is obtained, it is possible to locate fraud and fight against money laundering and aggressive fiscal engineering using the same algorithms as are used in social media. By uncovering this complex network of relationships, each taxpayer can be linked to his or her wealth and the path that connects them can be followed. In this way, the owners of a given network of companies and their relationships can also be identified, which facilitates the detection of tax fraud not only in direct taxes but also in value added tax (VAT).

The algorithms create, first of all, the network of relationships between individuals and companies. They then go on to discard those cases in which entrepreneurs set up a firm directly and those in which the corporate organisation has a real economic basis and pays tax. This provides the means to draw the map that leads from certain individuals to their hidden wealth in non-listed investee companies, infer their wealth, and so calculate whether fraud may have been committed through the use of these investee companies.

To date it has been possible, by means of complex statistical techniques, to carry out studies on income distribution, yet it has been almost impossible to conduct detailed person-by-person studies on wealth distribution. The research on which this article is based uses, for the first time, algorithms and artificial intelligence tools with the aim of describing the fabric of a company and the hundreds of economic, commercial, legal and personal relationships that exist within it.

The algorithms described here make it possible to gain extremely detailed knowledge of the possession of wealth, for the purpose of revealing fraud patterns. Ultimately, they constitute a mechanism that encourages the design of more effective policies for combating fraud, the creation of fairer and more precise tax policies, and the control of large fortunes.