One beneficial part of JPMorgan Chase's (NYSE: JPM) $2 billion trading fiasco is that it let us peer inside the too-big-to-fail banking world. We learned, for example, that Ina Drew, the JPMorgan executive overseeing the unit responsible for the losses, was paid an average of $1.3 million per month over the last two years -- a figure that made shareholders, most of whom had never heard of Drew, wonder who really owns these banks.

Why are bankers paid so much? Theories abound, but a speech (link opens PDF) last year by Andrew Haldane of the Bank of England made several enlightening points.

Haldane first points out how fast bank compensation has grown:

In 1989, the CEOs of the seven largest banks in the United States earned on average $2.8 million. That was almost 100 times the median US household income. By 2007, at the height of the boom, CEO compensation among the largest US banks had risen almost tenfold to $26 million. That was over 500 times the median US household income.

This boom didn't just enrich CEOs. In the 1950s, the average financial sector worker earned roughly the same wage as an average worker in all other industries. By 2007 the average finance worker outearned those in other industries by more than 50%:

Sources: Bureau of Economic Analysis and author's calculations.

Haldane has an interesting theory on why this happened. In short, he says we got the measurements all wrong when we started using return on equity as the most important metric of a bank's performance.

The problem is that return on equity can be goosed by taking on more risk through leverage, making a banker look really smart (and very rich) during a boom when all they're doing is loading up the balance sheet with debt. That leads to blowups when the economy hits a rough patch.

If a metric that did a better job adjusting for risk -- like return on assets -- were used, banker pay would be far more reasonable today. Haldane explains:

Imagine if the CEOs of the seven largest US banks had in 1989 agreed to index their salaries not to [return on equity], but to [return on assets]. By 2007, their compensation would not have grown tenfold. Instead it would have risen from $2.8 million to $3.4 million. Rather than rising to 500 times median US household income, it would have fallen to around 68 times.

Another variable banks have misguidedly rewarded pay for in recent years is size. My colleague Ilan Moscovitz and I once argued that banks have a history of paying managers for empire building even if it means lower performance than smaller peers. We used these two tables to make our point (which frankly made us laugh out loud when we made them):

JPMorgan Chase ($2 trillion in assets)

Year

Return on Assets

CEO Compensation

2005 0.7% $22.3 million
2006 1.1% $39.1 million
2007 1.1% $34.3 million
2008 0.2% $19.7 million
Bank of the Ozarks ($3.8 billion in assets)

Year

Return on Assets

CEO Compensation

20051.6%$464,997
20061.4%$774,064
20071.2%$825,588
20081.2%$912,336

Source: S&P Capital IQ.

"If you're a CEO, you don't build dynastic wealth for yourself by being small and nimble," we wrote. "You do it by being an enormous, sledgehammer-wielding giant."

The most common objection we encountered with this comparison was that it's harder to earn a high return on $2 trillion of assets than it is a few billion dollars of assets.

Our response is: Yes, that's exactly right. So why not break megabanks apart into smaller, more efficient pieces? People talk about breaking up too-big-to-fail banks because they endanger the financial system, but it could also be the best way to maximize shareholder returns. Just this year, renowned bank analyst Mike Mayo reckoned that JPMorgan's parts would be worth one-third more than the current market value if broken up. A similar case could almost certainly be made for Citigroup (NYSE: C), Goldman Sachs (NYSE: GS), or Bank of America (NYSE: BAC). Banks likely denounce these calls because executive compensation is increasingly tied to inefficiency (size) and risk (return on equity) which would fall if made smaller. Shareholder returns be damned; those yachts don't pay for themselves, you know.

Former Goldman partner Leon Cooperman said it best: "I determined many years ago that if you want to make money on Wall Street, you work there; you don't invest there. They just pay themselves too well."