How people get paid

The last of the roadmap's self-assessment targets, and the one where structure matters far more than the headline numbers — the numbers vary by firm, year, seat and cycle, and any specific figure dates almost immediately.

The four components

   BASE          modest by industry standards. Enough to live on;
                 rarely the point. Often capped well below the bonus.

   BONUS         the actual compensation. Annual, and either
                 discretionary (banks, most funds) or formulaic
                 (pod PMs — see below). Can be a large multiple
                 of base.

   SIGN-ON       paid on joining, to compensate for the DEFERRED
                 pay you forfeit by leaving your last employer.
                 Its existence is entirely a consequence of
                 deferral, and it is how firms poach.

   DEFERRED      a large fraction of the bonus, paid out over
                 3–5 years, contingent on still being there —
                 and sometimes on the firm's or your book's
                 later performance.

The rough shape: junior seats are base-dominated; senior seats are bonus-dominated with a substantial deferred component. At market makers organised as partnerships, senior compensation is a share of firm profits and can dwarf anything on this list — there are no outside shareholders to pay (M5.2).

Why deferral exists — the option problem

This is the part worth understanding properly, because it is a direct application of M2 and M3 rather than an HR convention.

Consider a trader paid a share of their annual profit, with no deferral and no clawback:

   good year   →  large bonus
   bad year    →  bonus of zero
   catastrophe →  bonus of zero, and you are fired

   your payoff is  max(profit, 0) × share

That is a call option on the firm's capital, granted free, and re-struck every January. And M2.3 told you exactly what a call is worth: it rises with volatility. The rational response to holding one is to take as much risk as you can get away with, because you keep the upside and the downside is bounded at zero.

Worse, from M3: the optimal way to maximise a one-year option payoff is a strategy with steady small gains and a rare enormous loss — sell tail risk, run carry, be short gamma (M4.6). Such a book looks superb for several years and then destroys far more than it ever earned. The individual is paid throughout, and the firm eats the final year.

Deferral is the fix. If most of this year's bonus only vests over the next three to five years, and can be reduced or cancelled if the book later blows up, then you are no longer holding a one-year option — you are holding something closer to a stake in the multi-year outcome.

   without deferral        payoff = a series of annual call options
                           → incentive to maximise volatility and
                             sell tails

   with deferral           payoff depends on surviving the horizon
                           → incentive to care about drawdown, which
                             is M3.6's point exactly

It is also, incidentally, a retention device — leaving means forfeiting unvested pay, which is why sign-on bonuses exist and why senior people are described as "golden handcuffed". Firms are happy for both effects.

Post-2008 regulation made deferral and clawback mandatory for material risk-takers at banks, precisely because the crisis was, in part, this incentive structure operating at scale.

The pod PM's deal

Pod shops are the exception to discretionary bonuses, and their arrangement is the clearest in the industry:

   a PM receives a DIRECT SHARE of their own book's P&L
   — reported figures commonly fall in the 10–20% range,
   negotiated individually and rising with track record —

   less:  the pod's own costs (team salaries, data, technology),
          which are typically charged against the book

   subject to:  the stop-loss (M5.3). A drawdown of roughly
                5–10% and the allocation, and often the job, is gone.

This is the purest incentive alignment in the industry and also the harshest. A good year can pay more than a decade elsewhere. A bad year ends the seat. It is why pod-shop turnover is high, why PMs are famously mercenary about moving, and why sign-on packages there are large — a firm poaching a PM must buy out both the deferred pay and the option value of an established allocation.

Note the interaction with pass-through fees (M5.3): the PM's team costs are charged to investors rather than absorbed by the firm. Paying more to retain a PM does not reduce the platform's margin.

Reading a compensation structure

The useful skill is not knowing the numbers — it is inferring what a structure is for:

   what you see                       what it tells you
   ─────────────────────────────────────────────────────────────────
   heavy deferral                     the firm fears tail risk and
                                      staff departure
   direct P&L share + stop-loss       a pod shop; you own your
                                      outcome, both ways
   partnership profit share           proprietary capital, no
                                      outside investors (M5.2)
   large sign-on offers               they are poaching, and buying
                                      out someone's deferred pay
   bonus mostly discretionary         the firm wants latitude to
                                      reward things P&L doesn't
                                      capture — or to pay less

And the general principle, which is the honest summary of all of M5:

Compensation structure is a risk-management tool, not just a cost. How a firm pays people determines what risks they take, so you can read the firm's understanding of its own dangers off the shape of its pay.

What M5 leaves you

You can now tell the three models apart and say why each is shaped as it is: a market maker running M1.7's equation on its own capital, long volatility and disclosing nothing; a pod shop renting capital to many uncorrelated teams under hard stop-losses, growing by hiring; a single-strategy fund with one edge and a capacity ceiling that often makes growth actively harmful. And you can read a pay structure as the risk model it is.

That is the whole of the roadmap's M5 self-assessment.

M6 is not a module with lessons. It is the reading habit that keeps all of this current — Money Stuff daily, one serious weekly, and the books for colour. The facts in M5 will drift; the structures won't, and the habit is how you notice which is which.

Source: for the option-on-the-firm argument, the post-crisis literature on banker compensation is extensive — the UK’s Walker Review and the FSB’s Principles for Sound Compensation Practices are the primary documents, and both are more readable than they sound. For current market pay, industry surveys and compensation consultancies publish annually; treat all specific numbers as perishable.