This is the year when bank shareholders started saying "enough" to gigantic CEO pay packages and giving a stinging thumbs-down to certain banks.

Paul Hodgson, a senior research associate at GMI Ratings, sums up this year's early proxy-season pay trends and offers insights into what's ahead.

A Briton, Hodgson has been researching and writing about executive pay for 16 years. He has authored a number of books, including "Building Value Through Compensation."

There's a widely held view that pay tied to short-term incentives contributed to the financial crisis. You wrote a year and a half ago that nothing had changed with pay practices since the Wall Street meltdown. Is that still the case?
PAUL HODGSON
: There hasn't been a great deal of change at many of largest banks we've had a chance to look at so far. Capital One has a performance share plan, although its CEO [Richard Fairbank] isn't part of it. But it's a very poorly designed plan that pays for below-median performance. Morgan Stanley has a similar plan, but its performance affects only a small part of long-term pay [of Chief Executive James Gorman and other executives]. Again it's poorly designed.

Goldman Sachs implemented a plan that measures performance over three years, based on its return on equity. But its ROE targets aren't particularly challenging. They [Goldman managers] may have known something we didn't at the time [the targets were established]. When an institution sets targets lower than they've achieved in recent years, it's a signal you need to heed. It might be that management knows it won't keep up its performance of recent years.

Among other big banks, Citigroup is the only one that made changes. It deferred a substantial portion of the cash bonus [for Chief Executive Vikram Pandit], and that only kicks in if Citi meets another performance target.

After Citi took Tarp [Troubled Asset Relief Program] money, and its share price got beaten to death in 2008, it took a bold move and granted some executives premium-priced stock options [options whose strike prices are higher than the market price at the time they're granted].

When you have a depressed stock price like Citi did, this is a very effective way of focusing executives on increasing the stock price to previous levels. With a premium price, you have to get stock up, and for it to stay up there, for like 30 days before you can exercise. Executives can goose it all they like, but it won't stay up for 30 days unless there's real value.

But when Citi gave stock options to Pandit in 2011, there was no evidence of a premium there.

Wells Fargo reacted to Tarp [pay restrictions] like a lot of other banks. It quintupled salaries and "reclassified" them as "salary stock." Then when it paid off Tarp, Wells reduced salaries, but not by very much. That's why so many of its named executives have outsized salaries.

Goldman Sachs took a slightly different route but ended up the same. It had been paying top executives $600,000 [in annual salaries] since its IPO and then suddenly started paying them $2 million. The $600,000 salaries were one of the things that Goldman had done right. If base salaries are low, the only way to make [big] money is through equity-based incentives that benefit from creating value.

What were the banks thinking in raising salaries so much?
They were taking advantage of a push by the Treasury and pay czar [Special Master for Tarp compensation Kenneth Feinberg] to increase salaries and have less pay at risk. And they were doing it in a fairly cynical way.

GMI has said that in addition to the size of CEO pay packages, it's important to look at how much larger they are than those of lower-ranking executives. On that relative basis, you've indicated that the pay of U.S. Bancorp CEO Richard Davis looks excessive.
More than half of S&P 500 companies have a differential [the difference in total pay between the CEO and that of other senior executives] that's more than three times. So it's widespread, but that doesn't make it right.

We have particular concern with banks. Their CEOs are not superstars but team leaders. If they're paid more than three times what anyone else is receiving, that doesn't sound like a team to me but one star and a couple of supporting actors.

That affects the morale of other senior executives and raises questions about succession planning. If you have a couple of CEOs waiting in wings, you'd want to pay them properly. It also affects the balance of power in the boardroom. A CEO who's paid lot more than anyone else can throw his weight around.

In the case of Discover Financial, the problem seems to be disclosure rather than dollars. Namely, that it doesn't let its owners know the details of how they're paying the hired help.
The SEC now requires disclosure as part of its regulations [release of the calculations that went into the previous year's pay package]. Some companies aren't comfortable disclosing future performance targets. But if the numbers are already in, everybody knows what the company's successes and failures were. Did it increase revenues? What were EPS?

Not disclosing targets set at the end of 2010 for 2011 seems to have no justification. Offering information that might aid competition isn't a problem because everybody knows how they've done. Instead, it could show that a company didn't set its [performance] targets high enough. Discover paid a cash bonus of almost $3.25 million.

So overall pay practices haven't changed much?
This [the wake of the financial crisis and passage of Dodd-Frank] was the moment when banks could have revolutionized compensation policy. Lots of advice was out there, but the vast majority have complied with regulation in a very nominal way and returned to business as usual.

What do you mean by "nominally compliant"?
They were asked whether their compensation policies encourage risk-taking. None said "Yes." It's difficult to take that at face value. Some do continue [to encourage inappropriate risk]. Look at non-U.S., mostly European, institutions. They made very substantial changes [in pay]. Deutsche Bank made changes. Credit Suisse made some pretty substantial changes. It increased the proportion of deferred incentives. It introduced two new deferred and future performance-linked plans, rather than just basing pay on annual performance. One measures ROE and share price growth over four years. In a deferred cash plan it uses ROE, which can increase or decrease cash bonuses. If you earn a cash bonus that bonus is deferred. But if ROE isn't steady or going up [over the next four years] then the eventual payout of that cash bonus is negatively affected.

The Dodd-Frank Act began requiring companies to hold say-on-pay votes among shareholders beginning last year, but this is the year they seem to be having a big effect. Why is that?
Part of the reason little changed last year is it [the say-on-pay vote] came about very quickly. I don't think shareholders were prepared for the analyses they had to make. There was huge dissatisfaction with ISS [Institutional Shareholder Services, the proxy advisory firm] recommendations. My sense is they were not as sophisticated as they needed to be so companies got a bit of a pass last year.

Are clawback provisions being implemented as you'd want?
Both Morgan Stanley and Goldman were involved in shareholder dialogues that caused them to give their clawback arrangements a bit more bite. My issue is that it's like closing the barn door after the horse has bolted. If these arrangements had been in place prior to the financial crisis, a lot of compensation that had been paid out at Wall Street banks would have been clawed back.

The SEC and other agencies are moving ahead with Dodd-Frank pay provisions that are supposed to regulate incentive pay at financial institutions. Where do we stand?
I haven't seen any final rules. The direction is similar to the regulations proposed in the U.K. and Europe. The other thing that's notable is that some of the largest banks are now subject to stress tests. Wouldn't it be a good idea for compensation committees to design incentives so that they have to pass the stress test before any incentives are paid out?