The Medici Lesson: Why Relationship Lending Survives the Algorithm

The Historical Precedent: Banking Before Algorithms

In 14th Century Florence, the Medici Bank dominated European finance. They had no credit ratings, no collateralized debt obligations, and no actuarial tables. What they had was relationship intelligence. When a wool merchant applied for a loan to ship English wool to Florence, the Medici agent in London physically inspected the wool. He knew its quality, its weight, and the reputation of the shipper. When the wool arrived in Florence, the agent there knew when it was sold, to whom, and at what price. The loan was secured not by a legal contract alone—though those existed—but by control of the physical flow of goods.

The Medici Bank survived wars, plagues, and the collapse of competing banks. It eventually failed only when management grew complacent and began lending to royalty (political risk) rather than merchants (commercial risk).

The lesson: Lending based on direct knowledge of the underlying asset and borrower has a 700-year track record of resilience. Lending based on models and ratings has a 50-year track record of periodic systemic collapse (1973 secondary banking crisis, 1998 LTCM, 2008 GFC, 2023 regional bank failures).

The Modern Parallel: The Rise of Private Credit

As of 2025, global private credit assets under management have surpassed i$1.7trillion.\; P$rivate credit now provides more than 60% of all leverage, up from 15% in 2008. The reason is simple: regional banks in the US and Europe have retreated due to Basel III endgame rules and higher capital requirements. In Asia, the retreat is even sharper—Hong Kong and Singapore have seen 14 consecutive quarters of declining bank lending to SMEs.

The hot topic is whether this growth is sustainable. Critics point to 2023’s distressed private credit funds (e.g., the UK’s PCF collapse) and warn of a 1998-style blowup. Proponents point to low historical defaults (sub-2% for senior secured private credit).

History says the truth is in the middle, but leans positive—provided the lender follows the Medici rules.

Four Medici Rules for Modern Private Credit

Rule 1: Control the collateral, not just the covenant

The Medici Bank didn’t simply take a lien on the wool. They had agents who could physically stop the wool from leaving the warehouse if the merchant missed a payment. Modern private credit often relies on “springing liens” or “blocked accounts” that activate only after default. By then, the collateral may have vanished.

We insist on direct operational control of three things for every credit position:

• Bank accounts: The portco’s primary operating accounts must be at a bank where we have a pledge agreement. No cash sweeps without our notice.

• Inventory: For warehouse-intensive businesses (logistics, commodities, manufacturing), we require weekly inventory reports and the right to send an auditor with 48 hours’ notice.

• Receivables: Factoring arrangements where we are the first-position beneficiary.

This sounds aggressive. But in a default scenario, these controls reduce recovery time from 18+ months (typical court-driven process) to 3–6 months. That 12-month difference is the difference between a 80% recovery and a 40% recovery.

Rule 2: Avoid “covenant-lite” at all costs

In 2006–2007, the syndicated loan market abandoned maintenance covenants. By 2008, defaulting borrowers had no obligation to maintain leverage ratios or interest coverage. Recovery rates fell from 70% (historical average) to 35%. Private credit has so far avoided this mistake, but pressure is building from institutional LPs who want higher yields without higher risk.

We track three non-negotiable covenants:

• Maximum senior leverage: 4.0x EBITDA, tested quarterly.

• Minimum fixed charge coverage: 1.25x.

• Minimum liquidity: 12 months of cash interest.

If a borrower won’t agree to these, we walk. There are always other deals.

Rule 3: Know the borrower’s borrower (the Medici network effect)

The Medici Bank knew not just the wool merchant, but also the wool merchant’s customers (the weavers) and suppliers (the shearers). That “second-degree” intelligence allowed them to see trouble coming before the merchant did.

Modern private credit can replicate this through vendor and customer checks. Before closing a credit facility, we interview the portco’s three largest customers and three largest suppliers. Off the record. We ask: “Are payments current? Is quality consistent? Are you looking for alternatives?” The answers often reveal more than the audited financials.

Rule 4: Float your own high-yield securitization only when rates are falling

In the late 1980s, Michael Milken’s Drexel Burnham pioneered the junk bond market. The winners were those who issued bonds in 1982–1984 (as rates fell from 15% to 11%). The losers were those who issued in 1987–1989 (as rates rose from 7% to 9%), triggering a wave of defaults when floating-rate debt repriced upward.

Private credit funds today face the same choice. Many are packaging their loans into CLOs (collateralized loan obligations) to recycle capital. Doing so in a falling-rate environment (like 2025–2026, as central banks ease) is smart: your underlying borrowers can refinance at lower rates, reducing default risk. Doing so in a rising-rate environment is dangerous. We time our CLO issuances for the back half of the cutting cycle.

The Hong Kong Advantage

Hong Kong’s private credit market benefits from three structural features:

1. No withholding tax on cross-border interest payments (unlike mainland China’s 10% or Singapore’s 15% in certain structures).

2. English common law for contract enforcement, which is familiar to international LPs.

3. The HKMA’s Account Opening Hub which has streamlined the KYC process for PE-owned portcos from 4 months to 6 weeks.

We are seeing a surge in cross-border private credit: a HK-based fund lends to a Vietnamese portco secured against its receivables from Japanese buyers, with disputes resolved in HK arbitration. Ten years ago, that deal would have been done by a Singapore bank. Today, the bank won’t touch it due to capital constraints. The credit fund earns 12–15% yield.

Algorithms cannot repo a shipping container in a typhoon. Credit ratings do not update when a supplier goes bankrupt on a Friday afternoon. The 2008 financial crisis proved that systemic risk grows when lending becomes anonymous and model-driven. The 2023 regional bank crisis proved that even regulated institutions can fail when their deposit base is digital and flighty.

Private credit, done well, is a return to the Medici model: slow, relationship-based, secured against hard assets, and controlled through covenants and operations. In Hong Kong, where trust is still the ultimate currency and reputation travels fast, this model is not a nostalgic throwback. It is the only rational response to a world where banks have retreated and algorithms have burned.