Here is my google doc

(pl click on the link; google sheet is given therein)

Screenshot is attached above

Deep-Dive Insights: Corporate Credit Evaluation Dataset

1. The Dataset's Hidden Story

Before diving into numbers, every row is a company knocking on a bank's door asking for credit. The bank then computed liquidity ratios, profitability metrics, and assigned a rating. This means every statistical pattern we find is really answering one question: "What separates creditworthy companies from rejected ones?" The numbers below are not just statistics , they are the financial fingerprints of corporate health.

2. Scatterplot Insights: The Geometry of Corporate Liquidity

Plot 1 : CurrentRatio vs QuickRatio (r = 0.2486) : "The Inventory Problem"

A correlation of only +0.2486 between two ratios that both measure short-term liquidity is striking. In theory, they should be closely related, but they aren't, and that gap tells a story.

The gap between CurrentRatio and QuickRatio = Inventory

A weak correlation means inventory levels are wildly inconsistent across companies in this dataset. Some companies are sitting on massive inventory (manufacturing firms, retailers) pushing their CurrentRatio high while QuickRatio stays low. Others hold almost no inventory (service firms, fintechs, banks) making both ratios nearly equal.

This is a red flag from a credit analyst's perspective : a company with a high CurrentRatio but low QuickRatio has liquidity that is trapped in unsold goods. If the economy turns, that inventory may not convert to cash fast enough to service debt.

Plot 2:  QuickRatio vs CashRatio (r = 0.7727): "The Strongest Signal in the Dataset"

This is the most reliable relationship across all three pairs, and it makes sense structurally:

Cash ⊂ Quick Assets ⊂ Current Assets (mathematically nested)

Since cash is a direct component of quick assets, they should co-move, and they do strongly. But r = 0.7727 (not 1.0) means some companies have high QuickRatios driven by large receivables rather than actual cash. Receivables are promises to pay; they carry counterparty risk.

From a credit evaluation standpoint, a company with high QuickRatio but low CashRatio is living off receivables, it looks liquid on paper but may face a cash crunch if clients delay payments. Banks should (and likely do) penalise such firms in the Rating column.

Plot 3: CurrentRatio vs CashRatio (r = −0.1989): "The Most Counterintuitive Finding"

A negative correlation between the broadest and narrowest liquidity measures is the most intellectually interesting result. Naively, you'd expect: more current assets → more cash. But the data says the opposite.

The explanation is a classic corporate finance trade-off:

Companies that aggressively accumulate current assets (inventory + receivables) are actually deploying their cash into working capital, leaving less in pure cash reserves.

This is the Working Capital Trap, fast-growing companies often have excellent CurrentRatios because they're scaling inventory and chasing receivables, but simultaneously draining their cash. From a lender's lens, these are the most deceptive borrowers, they look liquid but are operationally cash-starved. The negative covariance of −0.3671 confirms this is a real, systematic pattern across the 1000 firms.

3. Covariance: What the Numbers Actually Mean in Credit Context

Pair

Covariance

Financial Meaning

Cov(CR, QR) = +0.2828

Weak positive

Inventory differences dominate the relationship

Cov(QR, CashR) = +0.9552

Strong positive

Cash quality of liquidity is consistent

Cov(CR, CashR) = −0.3671

Negative

Working capital expansion crowds out cash

The asymmetry here is profound. QuickRatio sits in the middle, it covaries positively with both extremes (CurrentRatio and CashRatio), but CurrentRatio and CashRatio are inversely related to each other. This makes QuickRatio the most balanced and informative single metric for a credit analyst, it captures both the cash-richness and the working capital position simultaneously.

4. Mean & Standard Deviation: The Distribution of Corporate Health

The Negative Means are Telling

Both QuickRatio (μ = −0.8847) and CashRatio (μ = −0.8614) have negative means. In raw financial ratio terms, a negative liquidity ratio doesn't make conventional sense. This strongly suggests the data has been standardised or z-score normalised before being shared, meaning these aren't raw ratios but deviations from an industry benchmark.

This is actually more sophisticated, a negative QuickRatio here means the company is below the industry average in quick liquidity. A positive value means above average.

This reframes the entire analysis:

• μ(CurrentRatio) ≈ 0.025 → The average firm in this dataset is roughly at the industry benchmark for current liquidity, a balanced portfolio of applicants

• μ(QuickRatio) = −0.885 → The average applicant is significantly below industry benchmark in quick liquidity, this bank is seeing a lot of inventory-heavy, cash-light applicants

• μ(CashRatio) = −0.861 → Similarly below average on pure cash, reinforcing that these are operationally intensive firms with thin cash buffers

Standard Deviations: Who is the Wildcard?

Variable

σ

Implication

CurrentRatio

1.303

High dispersion, wildly different working capital structures

CashRatio

1.416

Highest, cash positions are the most unpredictable

QuickRatio

0.873

Lowest, most consistent metric across applicants

The fact that CashRatio has the highest standard deviation is alarming from a risk perspective. It means the bank is seeing companies ranging from extremely cash-rich to dangerously cash-poor. This bimodal-like spread in cash positions likely maps directly onto the Rating column — A-rated firms probably cluster at the high end, D-rated firms at the low end.

5. Chebyshev Bounds: The Risk Manager's Lens

Why Chebyshev Matters More Than Normal Distribution Here

Financial ratios are almost never normally distributed, they are right-skewed, fat-tailed, and sometimes bimodal (especially in a credit dataset where you have a mix of healthy and distressed firms). Chebyshev's inequality requires no distributional assumption, making it the honest statistician's tool.

Variable-by-Variable Credit Risk Interpretation

CurrentRatio (k=2): [−2.58, 2.63] At least 75% of applicant firms have a CurrentRatio within this band. Any firm outside (say, CurrentRatio > 2.63) might actually be hoarding unproductive current assets, which can signal poor capital allocation. Firms below −2.58 are dangerously illiquid and almost certainly in the Rejected and D-rated bucket.

QuickRatio (k=2): [−2.63, 0.86] Notice the asymmetric bound: the upper bound is only 0.86 (barely above average) while the lower bound is −2.63. This asymmetry reflects that most companies cluster in the negative (below benchmark) region. A company above 0.86 in QuickRatio is genuinely exceptional among this applicant pool, a strong Accept candidate.

CashRatio (k=2): [−3.69, 1.97] This is the widest interval among the three, spanning nearly 5.66 units. The sheer width of this band tells a credit officer: "Cash positions among our applicants are chaotic, you cannot rely on CashRatio alone to screen borrowers." This is why banks use composite scoring models (like the Rating column) rather than any single metric.

The 25% Outside the k=2 Band: Who Are They?

For all three variables, at most 25% of firms fall outside the k=2 Chebyshev band. In a dataset of 1,000 firms, that's potentially up to 250 firms that are statistical outliers on at least one liquidity dimension. These extreme firms are the most interesting credit cases, they represent either exceptional companies worth lending to at premium terms, or distressed firms that should be categorically rejected.

6. The Overarching Credit Narrative

Putting it all together, the dataset paints this picture:

The bank is evaluating 1,000 corporate loan applicants who are, on average, below industry benchmark on liquidity (negative means on QR and CashR). There is enormous variability in how firms achieve their liquidity positions, some through cash, some through receivables, some through inventory. The only truly reliable, internally consistent liquidity signal is the QuickRatio-CashRatio relationship (r = 0.77). CurrentRatio, despite being the most commonly cited metric, is the noisiest and most misleading signal due to inventory distortion. A sophisticated credit model for this bank should downweight CurrentRatio and focus on the QuickRatio-CashRatio pair as the core liquidity signal, supplemented by ROCE and P/E for profitability context.