Greetings,I am currently looking at bonus payments made to FTSE 100 directors. Strangely I find that in a cross sectional OLS regression that bonus seems to be negatively related to profit before tax. It a strange finding that baffles me and I know that the mistake must be in the modelling process but I am confined to the data set I have and OLS so was just wondering what you guys would do if this happened to you? Would you just state that while its statistically significant, it doesn't make much economic sense so ignore it or would you try and put a spin on it?Thanks in advance for any replies

Some things I would do:1. Make sure I haven't made a mistake.2. Are the time periods comparable; i.e. maybe the bonuses are paid in year n for performance in year n-1, but you are actually measuring profits in year n.3. Are the results statistically significant?4. Run an OLS of bonuses against the company's stated benchmarks (this might be something else, like revenue growth.)Finally, if all else fails, and 3. is true, you should be happy.You have a counter-intuitive result. Economists love counter-intuitive resultslike this, so you now have a publishable paper. regards, p.s. a last thought. Are you measuring the bonuses in a dimensionless way (say asa percentage of the total compensation, or some other denominator). Otherwise, itmay be apples vs. oranges, as larger companies would tend to pay more in absoluteterms no matter what.)

I am currently looking at bonus payments made to FTSE 100 directors. Strangely I find that in a cross sectional OLS regression that bonus seems to be negatively related to profit before tax.An economist would not find this result particularly strange.Read up on Agency Theory. This is basically the difference between what you pay people to do, and what they actually do.But even bog standard Black Scholes finance predicts this.A bonus is a sort of option, typically being based upon the "value" of the firm at year end. Very common for "bonuses" to be expressed directly as options for both tax reasons, and to make them more acceptable to shareholders.As we know greater variance of outcome pushes the value up.Consider a firm going through rough times, it's reasonable to expect a wider range of profits than a dull stable firm.Also bonuses for top level management are typically a function of 3 things:Share price, profit and achieving some specific goal.I note you've done profit, which correlates with share price of course, but is a quite different number. Also from what I've read most bonuses use this as the largest factor since of course the shareholders usually care more about this than a single year's dividend.A good wheeze is to set the bonus as a function of how you've performed compared to "similar firms", rather than the market as a whole. At one level this is quite reasonable, since if your profits went up even a little bit, whilst other firms in your sector all lost money, it's reasonable to say you've done a good job. But who are similar firms ? Should GM be compared to all other car makers, large car makers, large manufacturers, all manufacturers,all manufacturers with >50% of capacity in the USA ? The choice of comparative group can massively affect executive bonuses, and you can be sure that work is done on this."goals" can be all sorts of things, but a standard one is stuff like has happened at Sainsbury's where sorting out a horrible cost base took priority over annual profit.How did you deal with takeovers ? Frequently you find that these have specific extra incentives for executives, and M&A has been up, so you'd have to take that into account.You might be already seeing where the Agency Theory is coming in here.Manager get to set the strike price of these "options", and get to set their own goals, and frquently have insider knowledge on takeover activity, either as a predator or target.If a firm is doing badly, and the executives feel they will be ousted soon, then they will prefer bonuses over salary. Firstly setting a large % of their pay as "performance related" is good PR, but of course if they think they won't survive beyond the next bonus they won't care about salary very much.In theory the firm's auditors should catch this, but of course they have the personal integrity of accountants (yet more agency theory), and are appointed by the very same directors that are trying to make these schemes.Executive bonus/options are in any case really hard to price due to their complexity, and the fact that a firm of beancounters is not very likely to hire a top grade quant to value options in a way thtat pisses of the people that pay them.

Thanks for your replies. Alan, they are statistically significant and so I guess the problem lies within the model itself. The bonus for period n is paid for and accounted for in period n so I don't think that is the problem. DCFC, Agency theory is what I have been looking at. The twist is that I am trying to see if an indivdual can pick up a company annual report and from within there itself determine various types of information, agency theory being one. I understand that accounting based incentive schemes are actually quite useless as they do not fulfill many of the criteria various economists have outlined as to what the ideal incentive scheme should be so profit as per the company accounts might not be the best way to measure it.

MSC,My first guess was conflict of interest too. If the mgmt is profiting while the firm is doing badly, that's a sign. if this is indeed due to a conflict of interest (or Agency theory as dcfc says), you may have just stumbled across a good way to pick out the firms with a bleak future!

Economic Reality 101: Public companies are run for executive management, not shareholders.

QuoteOriginally posted by: MSCGreetings,I am currently looking at bonus payments made to FTSE 100 directors. Strangely I find that in a cross sectional OLS regression that bonus seems to be negatively related to profit before tax. It a strange finding that baffles me and I know that the mistake must be in the modelling process but I am confined to the data set I have and OLS so was just wondering what you guys would do if this happened to you? Would you just state that while its statistically significant, it doesn't make much economic sense so ignore it or would you try and put a spin on it?Thanks in advance for any repliesBonuses are part of the company's expenses. Therefore large bonuses can actually adversely affect profit before tax. You should compare bonus payments to profit before bonuses. Try recomputing under these circumstances. If it is still negatively correlated, then the company's bonus program is inefficient.You can also detect an agency problem this way. If pre-bonus earnings growth is commensurate to bonus growth, but not commensurate to final corporate profit growth, then the managers are getting the upper hand and eat up the shareholder's share of corporate profits.