Ownership structure

54 agreed on in connection with initial compensation grants. It is worth highlighting that the basic rationale behind both is similar: to tie compensation to factors that cannot be evaluated at the point in time of the grant.

The information on deferred compensation is incomplete as we lack observations for Société Générale and Unicredit in 2003 and 2007. However, considering the remaining dataset, we observe an uptick in the average deferred compensation as a share of the total compensation package. The average in 2003 and 2007 was 25,1% and 31,1%, respectively, and the 2011 figure is just below 50% (49,4%). The tendency is even stronger if we consider median figures, where the observations for 2003 and 2007 were 21,1% and 27,5% while the 2011 median is above 50% (51,3%). If we use the same panel in each of the years and disregard the observations we gathered for Société Générale and Unicredit also in 2011, the average share paid in deferred compensation is 51,2% and the median observation is 56,9%.

Considering clawback provisions, we observe a complete absence within the sample group compensation agreements in both 2003 and 2007. However, in 2011, 17 out of the 24 banks (71%) commented in their annual reports of SEC filings that part of the executive compensation package was subject to adherence to certain conduct and could be clawed back in cases where violation had taken place.

In brief, whether retroactive or deferred, our results give an indication that the long term focus has become more central in determining executive compensation. The variable cash compensation component is downplayed in favour of a stronger equity beta of the total compensation mix. Also, shareholder democracy has been extended to directly encompass executive compensation as a larger share of banks make the compensation committee suggestion conditional to shareholder approval via say-on-pay provisions. However, one issue with our results throughout the compensation data section is the high variability among the banks. While many factors have converged (reflected in a narrower max-min range), the spread remains significant.

5.4 Ownership structure

55 The ownership structure lays the foundation for determining the ability of investors to take concerted action to discipline a bank’s management team. It is one of the governance mechanisms that cannot be directly determined by the board of directors or the top executive team, and it in many instances, it is strictly external.

Defining a blockholder as a shareholder with 5% or more in direct or beneficial shareholdings in a company, we have mapped the number of blockholders and shares held in each of the three years. In both instances, we note a gradual increase between 2003 and 2011. On average, our sample had less than one large shareholder (0,7) in 2003, increasing to 1,1 in 2007 and more steeply to 2,0 in 2011.

The median increased from zero to one in 2007 and to two in 2011. The most dramatic increase was noted in Dexia where the number of blockholders increased from zero in 2003 to seven in 2011.

Excluding them from the sample group, the average in 2011 is 1,8 rather than 2,0.

Similar tendencies are reflected in the share of total capital controlled by blockholders. The average increased from 6,9% in 2003 and 12,3% in 2007 to 23,3% in 2011 and the median went from 0,0% to 8,2% in 2007 and 15,8% in 2011. Once again, Dexia is the extreme case with 82,2% of the equity controlled by blockholders in 2011 (0,0% in 2003 and 52,6% in 2007), but significant increases are noted in many cases. For example, neither Banco Santander nor Barclays, HSBC or Lloyds had and large shareholders prior to 2007. After the financial crisis, at least 12,0% of the equity was controlled by blockholders in each of the cases and for Lloyds, the number rose from zero to above 40%. The fourth UK bank in the sample, Royal Bank of Scotland, saw similar change: after having only one large shareholder in 2003 and 2007²³, controlling approximately 5% at both year-ends, the corresponding number increased to 66,9% in 2011 when the UK Treasury stepped in and saved the bank from imminent bankruptcy.

One type of shareholder – the government – is of particular interest from a stakeholder perspective.

Our mapping of government holdings in systemically important banks shows an interesting turn before and after the financial crisis. The only meaningful government involvement in 2003 and 2007 was recorded for Nordea where the Kingdom of Sweden held 19,0% and 29,1%, respectively at the two year-ends²⁴. In 2011, however, the situation was radically different. While most banks still did not disclose any government shareholdings, Dexia, Commerzbank, BNP Paribas, Unicredit, Lloyds and Royal Bank of Scotland saw the government stepping in to take significant equity stakes during the financial crisis. The two UK banks provide the most extreme cases with 66,9% and 40,3% respectively for RBS and Lloyds.

23 Santander in 2003 and Legal & General in 2007.

24 A minor stake of 0,6% in Credit Suisse was held by the Suisse government in 2007.

56 The disclosed shareholdings among directors show that insiders rarely hold a significant stake in large financial institutions. The average share of common equity held by insiders was below 1% in each of the years, and the trend is negative if anything, declining from 1,0% in 2003 to 0,8% in 2007 and 0,6%

in 2011. Most banks do not have insiders who, as an aggregate group, hold more than 1% and the one large exception from the general pattern is Société Générale where insiders controlled 8,5% in 2003 and 7,2% in 2007 and 2011²⁵. If we exclude them from the sample, the averages for the three year ends fall to 0,7%, 0,5% and 0,3%, respectively.

25 The shares are held by employees and former employees via an employee share ownership program (Société Générale 2011, p 25).

57