Meridian Devices: $38 Million Trapped and $15 Million Borrowed at the Same Time
Meridian Devices has $96 million of consolidated cash and a $250,000 target operating balance for each of its 63 accounts. Brazilian, Chinese and Indian subsidiaries hold $38 million that cannot be repatriated as dividends this year: Brazil imposes a minimum local balance and 15% withholding tax on dividends, China requires SAFE approval for outbound intercompany loans, and India caps interest rates on intercompany loans. Meanwhile the US parent has $15 million drawn on its revolver at 8.2%, and European entities hold $22 million on deposit earning an average of 4.0%. Treasury builds the daily cash position from prior-day bank statements in a spreadsheet and forecasts weekly; forecast variance averages 22%. The CFO asks whether building an in-house bank is the right response.
1. Quantify the cost of the current misallocation and assess whether an in-house bank actually solves it.
Read the worked answer
Quantify the two sides.
- Borrowing cost: $15M x 8.2% = $1.230M/year.
- Interest earned: $22M x 4.0% = $880k/year.
- The offsettable spread is $15M x (8.2% - 4.0%) = $630k/year, not the full $1.23M, because only the borrowed amount can be repaid from the deposits.
Separate movable from trapped cash. Of the $22M in Europe, assume ~$10M can be swept or lent to the parent without friction (local regulatory capital, cash-pool minimums, and FX may block the rest). Repaying $10M of revolver with $10M of 4.0% deposits saves 10M x 4.2% = $420k/year before tax and transaction costs - real, but modest.
An in-house bank cannot repatriate restricted cash. China's SAFE approval and Brazil's withholding tax are legal constraints, not structural ones. What an IHB can do is (a) pool where permitted, (b) net intercompany flows multilaterally so gross funding needs shrink, (c) lead/lag intercompany pricing within transfer-pricing rules, and (d) invest trapped balances locally at better rates.
Judgement: a full IHB costs roughly $300-500k to build plus staff and governance. Against ~$420k of annual savings it is marginal on this arithmetic alone. Recommend a phased approach: fix the forecasting and pooling fundamentals first, run a payments-on-behalf-of structure where legally feasible, and revisit the IHB when volumes and trapped-cash strategies justify it. The strongest candidates will say the IHB is a tool, not a diagnosis.
2. The 22% forecast variance is the larger problem. What should treasury do, and how does variance drive the buffer and revolver size?
Read the worked answer
Treat variance as the root cause. The parent is borrowing $15M while holding $22M because the position is built on prior-day statements and a weekly forecast with 22% error. Rebuild the forecast bottom-up: 13-week direct forecast by entity and by driver (payroll dates, tax installments, top-50 customer collections, debt service), reconciled monthly against actuals, with variance categorized as timing, amount, or entity error.
Convert variance into a buffer requirement. If average weekly net outflow is $18M with a standard deviation of $4M, a 95% buffer adds roughly 1.65 x $4M = $6.6M of precautionary liquidity. That buffer is either idle cash (cost: the spread) or revolver availability (cost: the commitment fee). Reducing variance from 22% to 10% roughly halves the buffer - a larger and more durable benefit than sweeping a few million of deposits.
Practical steps:
- Set target balances per account from transaction volume and bank minimums, not a flat $250k across 63 accounts; close or zero-balance the dormant ones.
- Move to same-day/intraday bank reporting rather than prior-day statements.
- Institute a daily position with a documented escalation threshold (e.g., variance > $2M triggers a call).
- Size the revolver against the forecast distribution, not the worst historical week.
A good answer notes that forecast accuracy is a treasury KPI in its own right and should be reported to the CFO monthly.
3. Evaluate notional pooling, physical pooling, and intercompany lending for the European entities. What breaks each one?
Read the worked answer
Physical pooling (zero-balance accounts): sweeps each entity's balance to a header account daily. Simple, available almost everywhere, and preserves legal ownership of cash. Downsides: creates intercompany balances that must carry arm's-length interest, generates gross intra-group flows and FX conversions, and strips local entities of visible cash (a behavioural and audit issue). Best default when the entity set is stable and currencies are few.
Notional pooling: the bank offsets balances without moving cash, preserving local ownership and balance-sheet position. It fails in practice because: the bank must take cross-jurisdictional credit risk and usually demands cross-guarantees; offset rights must survive insolvency in each jurisdiction (not true in many countries); it generally requires a single currency (EUR pooling is viable where local law allows); and pricing is opaque.
Intercompany lending: flexible and cheap internally, but constrained by withholding tax on interest, thin-capitalization and interest-deduction limits, transfer-pricing documentation, and FX/regulatory approvals (SAFE in China, interest caps in India).
Judgement: choose pooling by jurisdiction rather than one global model - physical pooling for the Eurozone, no pooling for China/India where repatriation is the constraint, and invest trapped balances locally at term rather than forcing them into a structure that only looks efficient on a consolidating spreadsheet.