Investment Banks: The Complete Guide

An investment bank is a financial institution that helps corporations, governments, and large investors raise capital, advise on transactions, trade securities, manage assets, and provide research and brokerage services. Crucially, it does not take retail deposits and does not (in its pure form) make loans to consumers. That is a commercial bank's job.

The clearest way to understand an investment bank is to look at who its clients are. They are not individuals. They are corporations issuing stock or debt, governments selling sovereign bonds, private equity firms buying and selling companies, hedge funds trading and borrowing, sovereign wealth funds and pension plans deploying billions, and ultra-high-net-worth families using private wealth services. The minimum size of a meaningful relationship typically runs in the hundreds of millions or billions of dollars.

Until 1999 in the United States, the Glass-Steagall Act formally separated commercial banking from investment banking. The Gramm-Leach-Bliley Act repealed that separation, allowing the universal bank model that JPMorgan Chase, Bank of America, and Citigroup operate today: one corporate parent with a commercial bank, an investment bank, an asset manager, and consumer-facing operations all under one roof. Goldman Sachs and Morgan Stanley converted to bank holding companies during the 2008 financial crisis to access Federal Reserve liquidity but remain primarily investment-banking-focused. Pure stand-alone investment banks (the old Lehman, Bear Stearns, Merrill Lynch model) have effectively disappeared.

An investment bank is also distinct from an asset manager (which holds and invests other people's money under a mandate, like BlackRock or Fidelity), a hedge fund (which manages money for sophisticated investors and trades for absolute return), and a private equity firm (which buys whole companies). Investment banks routinely interact with all three but do not do their primary job.

The investment banking industry is tiered. At the top, the bulge bracket consists of the largest, most diversified global firms that serve the biggest clients on the biggest deals. The traditional bulge bracket today is Goldman Sachs, Morgan Stanley, JPMorgan Chase, Bank of America, Citi, and depending on who you ask, Barclays, Deutsche Bank, UBS, and Credit Suisse (the last now absorbed into UBS after the 2023 emergency takeover).

Below them, elite boutiques are advisory-focused firms that punch far above their weight in M&A and restructuring. Lazard, Evercore, Centerview, Moelis, Rothschild, PJT Partners, Houlihan Lokey, and Perella Weinberg sit in this tier. They have no balance sheet to speak of, no trading floors, and no underwriting. Just advisory. They compete with the bulge bracket on the largest deals because their senior bankers, almost all former bulge-bracket MDs, bring deep client relationships and judgment without the conflicts of interest that come with a balance sheet.

The middle market covers Jefferies, Stifel, Piper Sandler, Raymond James, William Blair, Baird, and similar firms. They serve mid-sized companies (broadly $50M to $1B enterprise value) that are too small for the bulge bracket to focus on but large enough to need full-service advisory and capital-raising. Jefferies in particular has grown aggressively and now competes with the bulge bracket on a meaningful share of deals.

League tables — published by Bloomberg, Refinitiv, and Dealogic — rank banks by deal volume in M&A advisory, equity underwriting, and debt underwriting. Senior management cares about league-table position because it drives client perception and the fee pool. The honest truth is that league tables are partly an arms race: banks negotiate hard for credits on deals where they did marginal work, because relative position matters.

Outside the global tiers, every major financial centre has regional specialists. Standard Chartered and Investec in emerging markets. Renaissance Capital in Africa, the CIS, and emerging markets. Numis (acquired by Deutsche Bank in 2023) historically in UK small and mid caps. EFG Hermes in MENA. Stanbic and Renaissance in sub-Saharan Africa. The economics differ by geography but the structure rhymes.

The Investment Banking Division — IBD or banking — is the advisory and capital-raising business. It is the part most people picture when they hear 'investment bank.' Its work falls into two main categories: corporate finance advisory (M&A and restructuring) and capital raising (equity and debt issuance).

IBD is organised along two axes. Coverage groups are organised by industry — Healthcare, TMT (Technology, Media, Telecom), FIG (Financial Institutions Group), Consumer & Retail, Industrials, Natural Resources, Real Estate, Power & Utilities, Sports & Entertainment. Coverage bankers own the client relationship for any company in their sector. They know the CEO and CFO. They know the strategic situation. They are the ones who get the call when the company is thinking about a deal. Product groups, by contrast, are organised by deal type — M&A, ECM, DCM, Leveraged Finance, Restructuring, Private Capital Markets. Product bankers are the technicians who execute the specific transaction once a coverage banker has won the mandate.

On any given deal, a coverage team and one or more product teams work together. A healthcare-coverage MD might bring a deal to a healthcare company's CFO; the M&A product team and the ECM product team then execute the transaction. Coverage owns the client. Product owns the trade.

The hierarchy in IBD is rigid. Analysts (typically two-year rotations directly out of undergrad, working 80-100 hour weeks on financial models, presentations, and processing) are the most junior. Associates (post-MBA hires or analysts promoted internally) own the work product and manage the analysts. VPs (typically 3-4 years post-MBA) manage process and run day-to-day execution on deals. Directors and MDs are the senior people who originate deals, manage client relationships, and price the bank's services. The journey from analyst to MD typically takes 12-15 years.

A typical week for an analyst is dominated by 'pitch books' — slide decks of analysis the bank uses to win business — and live deal work. Models include three-statement operating models, M&A merger models with accretion/dilution analysis, leveraged buyout models, and discounted cash flow models. The work can be brutal in volume. The intellectual challenge is uneven; the volume is constant. The compensation reflects this: in 2024-2025, first-year analyst total compensation at major US banks ranged roughly $175,000 to $225,000.

M&A is the highest-margin business in investment banking by a wide margin. The fee on a $5 billion sell-side advisory mandate can run $40-80 million for the lead advisor. The cost of execution — a small team, a few months — is tiny relative to the fee. This is why every bank fights for M&A market share, and why elite boutiques have built billion-dollar businesses on advisory alone.

M&A divides into sell-side and buy-side mandates. On the sell-side, the bank represents a company that is putting itself up for sale (or a division that the parent wants to divest). The work begins with valuation, then preparation of marketing materials (a Confidential Information Memorandum, the 'CIM'), then a controlled auction process with potential buyers (typically other strategics in the industry and financial sponsors / private equity), then negotiation of price and terms, then closing. The fee is contingent on a deal closing — typically 0.5% to 1.5% of transaction value, scaled to the size of the deal.

On the buy-side, the bank represents an acquirer. The work is similar in structure but the dynamics differ: the buy-side advisor helps identify targets, run financial analysis, structure financing, and negotiate. Buy-side fees are usually smaller (often a fixed fee plus a smaller success component) because the buyer typically shops the deal across multiple potential advisors before committing.

A specific deliverable that almost always falls to the M&A bank is the fairness opinion. This is a written legal opinion, signed by the bank, that the consideration paid (or received) in a transaction is fair from a financial point of view to the relevant party. It is what the company's board uses to defend the transaction in the event of a shareholder lawsuit. Fairness opinions are produced by a separate fairness committee inside the bank that does not work on the deal day-to-day. They typically come with a substantial fee in their own right.

Specific transaction structures matter. A merger of equals, an asset sale, a stock-for-stock acquisition, a leveraged buyout, a tender offer, a SPAC merger, a divestiture, a spin-off — each has different tax consequences, different accounting treatment, different regulatory hurdles, and different financial implications for both sides. Senior bankers know these distinctions cold, and clients pay them in part for that knowledge.

Accretion / dilution analysis is the standard test of any stock-funded M&A deal: does the transaction increase or decrease the buyer's earnings per share, both pro forma year one and after expected synergies? It is one of the simplest analyses in finance to execute and one of the most heavily fought-over in negotiation. The analysis is often the difference between a deal getting done and not.

Restructuring is M&A's distressed cousin. The bank advises companies (and sometimes their creditors) on Chapter 11 reorganisations, out-of-court restructurings, debt-for-equity swaps, and exchange offers. Houlihan Lokey, Lazard, PJT, and Evercore are the dominant names. Restructuring is counter-cyclical: when M&A volume falls in a recession, restructuring volume rises.

ECM is the business of raising equity capital for clients. The flagship product is the initial public offering: a private company sells stock to public investors for the first time. ECM also handles follow-on offerings (additional equity issued by an already-public company), at-the-market offerings (continuous selling at prevailing market prices), convertible bonds (debt that can convert into equity), and rights issues (where existing shareholders are offered new shares pro rata).

An IPO is a long process — typically 4-9 months from kickoff to first trade. It begins with the bank pitching to win the lead-bookrunner role. Once mandated, the bank organises the diligence process, drafts the prospectus jointly with the company's lawyers (the 'red herring' preliminary version, then the final filed version), conducts management roadshows with institutional investors, builds the order book, prices the offering on the night of the deal, and allocates shares the next morning. The first trade typically happens the day after pricing.

Pricing is where the bank's reputation and judgment matter most. Price the deal too high and it 'breaks issue' — trading below the offer price on the first day, embarrassing the company and the bank. Price it too low and the company leaves money on the table. The bank's job is to find the highest price at which the offering will be 'covered' — that is, where institutional demand exceeds supply by enough to support the stock in early aftermarket trading. Long-running rules of thumb suggest 5-15% first-day pop is healthy. Anything more starts to look like the bank deliberately underpriced to favour its institutional clients.

The fees on an IPO — the gross spread — are typically 5-7% of proceeds for US deals, paid by the issuer and split among the underwriting syndicate. The lead bookrunner takes the largest share, around 20-25% of the gross spread, plus a Praecipuum (a separate fee) for running the books. Active joint bookrunners take 10-15% each. Co-managers take a few percent. The remaining 30-40% is the underwriting allowance and selling concession, distributed in proportion to actual selling done.

Lock-ups are nearly universal: insiders and pre-IPO holders agree not to sell their shares for 90-180 days after the IPO. This prevents an immediate flood of supply that would crater the stock. The lock-up expiration is itself a market event — the first trading day after lock-up release is often weak as previously locked-up shares hit the market.

Follow-ons are simpler. A company files a registration statement, the bank takes a few days to a few weeks to market and price the offering, and proceeds typically take a small discount (3-7%) to the prevailing market price. ATM offerings are even simpler: the company files a shelf and dribbles stock into the market at prevailing prices over months or years. Convertibles are technically structured products: the bank prices them as a combination of a bond and a call option, and they appeal to issuers because the embedded option value reduces the explicit interest cost.

DCM raises debt capital for the same kinds of clients. The most common products are investment-grade corporate bonds (issued by BBB- and above-rated companies, the bulk of the volume), high-yield bonds (BB+ and below, the speculative-grade tier), syndicated loans (large bank loans split among multiple lenders), leveraged loans (the loan equivalent of high-yield bonds, often used in LBO financings), structured credit (asset-backed securities, mortgage-backed securities, CLOs), and sovereign debt for governments.

DCM is a higher-volume, lower-margin business than ECM. Investment-grade bond underwriting fees are typically 0.35-0.875% of proceeds, depending on tenor and complexity. High-yield fees run 1.25-2.0%. The execution timeline is much faster — a routine investment-grade bond can be priced and allocated in a single day, sometimes in hours. There are no roadshows for vanilla IG bonds; the issuer announces the deal in the morning, the bank builds the book during the day, and pricing happens that afternoon.

Syndicated and leveraged loans live in a slightly different world. They have a documentation phase (the credit agreement), a syndication phase (selling participations to other banks and institutional investors), and an allocation phase. The bank that arranges the loan earns an arrangement fee plus the rate spread on its retained portion. Leveraged loans in particular have grown into a $1.5+ trillion market in the US alone, with a deep institutional investor base of CLOs, loan mutual funds, and direct lenders.

DCM's economics are driven by relationships and balance-sheet capacity. To win a bond mandate, a bank usually needs to have lent to the issuer at some point — corporates reward banks that provide credit with the better-paying capital-markets business when they next issue debt. This is why universal banks (JPMorgan, BofA, Citi) tend to dominate IG league tables: they have the largest corporate-lending books, which gives them the biggest pool of bond mandates to compete for.

Sales and trading — S&T or 'the floor' — is the business of buying and selling securities and derivatives for the bank's institutional clients. The two big silos are equities and FICC (Fixed Income, Currencies, Commodities). Within each, sub-desks specialise by product: cash equities, equity derivatives, prime services, and program trading on the equity side; rates, credit, FX, emerging markets, and commodities on the FICC side.

The sales side is client-facing. Sales coverage maintains relationships with institutional buyers and sellers — hedge funds, mutual funds, pension plans, insurance companies, sovereign wealth funds. Salespeople pitch trade ideas, take orders, market new issues, and feed flow information back to the trading desk. They are paid on production credits — a complex internal system that allocates revenue to the people who originated each trade.

Trading is the book-runner. Traders make markets — quoting bid and ask prices that clients can hit or lift — and manage the risk of the resulting positions. A market-making desk earns the bid-ask spread on flow and tries to hedge or unwind inventory before the market moves. Profitable market making requires deep knowledge of order flow, fast risk management, and the willingness to take inventory positions clients want to offload.

Before the 2010 Volcker Rule (a Dodd-Frank provision implementing a ban on US bank proprietary trading), large parts of S&T were proprietary — traders deployed the bank's own capital to take outright positions on directional and relative-value views. The Volcker Rule largely ended that, with carve-outs for market making and underwriting. Banks have continued to take positions disguised as 'inventory' but the scale has come down dramatically. Genuine prop trading now lives at hedge funds and at firms like Citadel Securities, Jane Street, and Hudson River Trading, which have eaten much of the market-making business banks used to dominate.

S&T compensation is heavily formulaic. A senior trader on a profitable desk can earn $5-15M+ in a strong year; the same trader on a losing desk can be cut at year-end. The pay-for-performance structure is sharper than IBD, and the variance is correspondingly higher.

FICC is much larger than equities by revenue. The FX market alone trades $7+ trillion a day; rates and credit add multiples of that. The largest banks (JPM, Goldman, Citi, BofA, Morgan Stanley, Barclays, Deutsche, BNP) all run global FICC franchises, though the post-2008 era has seen consolidation, with European banks pulling back and the top US banks gaining share.

Sell-side equity research analysts cover individual stocks, publish reports with rating recommendations (Buy, Hold, Sell), and host investor calls. Their primary clients are institutional fund managers — both at hedge funds and at long-only mutual funds.

The 'Chinese wall' is a regulatory and ethical barrier between research and the rest of the bank, especially IBD. Research is supposed to publish without IBD influence; IBD is supposed to use research only after publication. Pre-2003, this wall was famously porous; analysts at major banks were privately disparaging stocks they were publicly recommending to favour IBD relationships. The 2003 Global Settlement (Spitzer) imposed structural separations and fines totalling $1.4B across ten banks. The wall is now genuinely meaningful, although it leaks at the margins through deal-specific carve-outs (banker-research-investor calls during IPOs, etc).

MiFID II (in effect 2018) was the next regulatory shift. It required European institutional clients to pay separately for research instead of bundling it into trading commissions. The unbundling has compressed research budgets dramatically. Many sell-side research departments have shrunk by 30-50% since 2018; some have been spun out as standalone businesses; some smaller stocks have lost coverage entirely.

What an analyst publishes: initiating-coverage reports (deep, often 50-100 pages, when picking up a new name), models updated through earnings, quarterly previews and reviews, industry / thematic notes, and event-driven flash notes. Senior analysts also organise investor field trips, host management at conferences, and publish stock-picking lists for their clients. The buy ratings are mostly read; the sell ratings are mostly ignored. The hold ratings are mostly noise.

Research analysts are now compensated mostly out of the trading commission pool and out of investor-vote 'broker reviews' (institutional clients' annual ranking of analysts they value most). Compensation for senior analysts ($500K-2M for managing directors at major firms) is below what comparable seniority earns in IBD or trading. The reason people stay: the work is genuinely intellectually engaging, the lifestyle is better, and the buy-side hedge-fund exit option is real and lucrative.

Asset management and wealth management are the two parts of the bank that hold and invest other people's money under a mandate. They are different businesses, though they are often linked.

Asset management (AM) serves institutional clients (pension funds, sovereign wealth funds, insurance companies, large endowments) and pooled retail vehicles (mutual funds, ETFs, separately managed accounts). The product is a fund or strategy with a stated investment objective; the fee is a percentage of assets under management (AUM) plus, for some products, performance fees. Goldman Sachs Asset Management, JPMorgan Asset Management, BlackRock (the world's largest, $11T+ AUM), Morgan Stanley Investment Management, BNY Mellon, Northern Trust, State Street are the major names. The economics are scale-driven: a fund that has 10x more AUM than its competitor at the same fee rate has 10x more revenue.

Wealth management (WM) serves wealthy individuals and family offices. The break-points vary, but typically: mass affluent ($100K-1M, served by mass-market platforms), high-net-worth ($1-30M, served by private bankers), ultra-high-net-worth ($30M+, served by dedicated client teams). The product is comprehensive financial advice — investment management, lending, estate planning, tax structuring, philanthropic advice. The fee is usually a percentage of investable assets (1.0-1.25% for HNW, lower for UHNW) plus product fees.

Morgan Stanley's wealth management business is the post-2008 success story of the industry. James Gorman built it from the Smith Barney acquisition (2009) into a $5T+ AUM franchise that is now the largest profit centre in the firm — bigger than IBD or S&T. JPM, Bank of America (Merrill Lynch), Goldman Sachs (the smallest of the big four), and UBS are the other major US-anchored wealth managers. Globally, UBS (post Credit Suisse absorption) is now the largest by HNW AUM.

Both AM and WM have benefited from the long bull market in financial assets and from the index/ETF revolution. They are also the cleanest businesses on the income statement: predictable fee revenue, low capital intensity, scalable margins. Every bulge-bracket bank now wants more of both.

Prime brokerage is the part of an investment bank that services hedge funds. It provides custody (holding the fund's securities), securities lending (lending out the fund's long positions and lending in securities the fund wants to short), financing (margin loans against the fund's positions), capital introduction (introducing the fund to potential institutional investors), and operational support.

The economics are large but tight: prime brokerage runs on net interest margin (the spread between what the bank pays to borrow and what it charges hedge fund clients to finance their positions) and securities-lending fees (the rebate the bank keeps on stock loans). A hedge fund client with $5B in assets and 3x gross leverage is generating millions in annual revenue for its primes through interest, financing, and stock-loan fees.

Goldman Sachs and Morgan Stanley have historically been the two largest prime brokers, each with roughly 25-30% market share. JPMorgan, Barclays, and Bank of America make up the rest of the top tier. Hedge funds typically use 2-3 primes simultaneously — concentrating with one bank creates concentration risk, but spreading too thin loses access to the best service tier.

The Archegos collapse in March 2021 was the cautionary lesson. Bill Hwang's family office had built $30B+ in synthetic equity exposure through total-return swaps with multiple prime brokers. When margin calls hit, the unwind cost the banks $10B+ in aggregate losses — Credit Suisse $5.5B, Nomura $2.9B, Morgan Stanley $911M. Goldman and JPMorgan unwound their positions faster and largely escaped. The episode revealed that even with apparently sophisticated risk management, banks can be carrying enormous concentrated exposure to a single client across multiple desks without anyone connecting the dots. Every major prime broker rebuilt its risk frameworks afterwards.

The investment banking career path is one of the most rigid in professional services. The titles, the years-in-grade, the responsibilities, and the comp progression are remarkably standardised across firms.

Analyst (years 0-2). Direct-from-undergraduate hires. The work is execution: building models, formatting decks, processing diligence, supporting deal teams. Hours run 80-100 per week, often longer near deal close. Total compensation in 2024-2025: first-year ~$175-225K (salary $110K base, bonus $65-115K), second-year ~$200-260K. After two years, most analysts either get promoted internally to associate (about 30% at major banks) or leave for private equity, hedge funds, or business school.

Associate (years 2-6). Hired post-MBA or promoted from analyst. The work shifts to managing the analyst, owning the model, and coordinating with the VP. Hours moderate slightly to 70-90/week. Total comp: first-year associate $300-400K; senior associate $400-550K. Associates begin to develop client-facing skills but do not yet originate.

Vice President (years 6-9). The first level where you start running deals end-to-end and managing junior bankers. Hours are 60-80/week and the work is more cerebral: structuring deals, managing client relationships at the working-team level, coordinating product groups. Total comp: $600K-1M.

Director / Executive Director (years 9-12). The level just below MD. By now you should be originating some business or playing a meaningful client-relationship role. Comp: $800K-1.5M. The pyramid narrows aggressively here — the survival rate from VP to MD is significantly less than 50%.

Managing Director (year 12+). The level that originates business, prices the bank's services, owns client relationships, and gets paid for it. Total comp ranges enormously: a junior MD at a bulge bracket is earning $1-3M; a senior MD running a major coverage group or product group can earn $5-15M+; the very top heads of investment banking divisions earn $20M+. The MD level is also where firing risk increases: MDs who don't bring in business get cut, often quickly.

Exit options matter as much as the ladder. From analyst, the most-trodden paths are private equity (the 'on-cycle' recruiting process happens in the analyst's first year, with offers extending to start two years later — yes, 18 months before the actual job begins), hedge funds, business school, corporate development at a strategic, growth equity, family offices, and operating roles at portfolio companies. From associate or VP, the same exits exist but the bar is higher and the deal flow is competitive. From MD, the natural moves are to run a smaller bank, lead an advisory boutique, or move into private equity at a senior level. Exits to startups and tech are common at every level.

An investment bank's revenue breaks down into three buckets. Fee-based income comes from advisory mandates (M&A, restructuring) and underwriting (ECM, DCM). It is high-margin but lumpy and procyclical: surges in good times and craters in bad. Spread / flow income comes from sales and trading market-making and prime brokerage financing. It is more stable than fees but lower-margin and balance-sheet-intensive. Asset-management and wealth-management fees are the most stable: a percentage of AUM that grows roughly with the market.

Different banks have different revenue mixes. Goldman Sachs and Morgan Stanley historically derived 50%+ of revenue from trading and 30-40% from IBD; both have shifted aggressively toward asset and wealth management since 2010. JPMorgan's investment bank is one of four major segments alongside consumer banking, commercial banking, and asset/wealth management — meaning the parent-level revenue is dramatically more diversified. Pure boutiques like Lazard and Evercore are nearly 100% advisory.

Regulatory capital is the binding constraint on every bank that has a balance sheet. Under Basel III and IV (the latter being phased in through 2030 in most jurisdictions), banks must hold capital against their risk-weighted assets. Trading and lending positions consume risk-weighted assets and therefore capital; pure advisory does not. The post-2008 regulatory tightening is a major reason Lazard and Evercore have grown profitable: they have no balance sheet, no capital constraints, and can operate at higher returns on equity than full-service rivals.

Return on equity is the metric senior management is judged on. Pre-2008, large investment banks routinely earned 20-30% ROE on heavy leverage. Post-Basel III, leverage came down hard and ROEs initially fell to 5-10%. By 2024, the best-run firms (JPMorgan, Morgan Stanley) were back to 14-17% ROE, but the industry-wide pre-2008 returns are gone forever. The shift toward asset and wealth management is partly a response to this: those businesses produce 25-30%+ ROEs at scale and don't consume regulatory capital.

The mistakes that destroy investment banks tend to rhyme. Excessive leverage applied to bad collateral is the canonical pattern: Long-Term Capital Management in 1998 (a hedge fund, but it nearly took down its prime brokers), Bear Stearns in 2008 (overleveraged with mortgage paper), Lehman Brothers in 2008 (same pattern, larger), Credit Suisse in 2022-2023 (a long string of risk-management failures: Greensill, Archegos, the Mozambique tuna bonds, the spy-on-employees scandal, and a deposit run that finished it). Bear and Lehman fell because of mortgage exposure, but the underlying disease was always the same: too much leverage, too little capital, and a culture that failed to enforce risk discipline at the top.

What does a healthy investment bank look like in 2025? Diversified revenue (no single business above 40%), high-quality fee mix (M&A and AM/WM rising; pure trading flat or down), well-capitalised balance sheet (CET1 ratios well above the regulatory minimum), and a rigorous risk culture. The four US banks meeting all four criteria today — JPMorgan, Goldman Sachs, Morgan Stanley, and Bank of America — are why the industry has stabilised. The European banks remain in a slow restructuring out of the businesses where they cannot compete and into the ones where they still can.