Almost every form of foreign money leaves a country when it gets into trouble. Portfolio investors sell, direct investment pauses, and lenders reprice risk. Remittances do the opposite: they rise. Understanding why matters, because that counter-cyclicality is the single most valuable property of the flow — and it is routinely explained in ways that are either sentimental or wrong.
This piece works through the actual mechanisms. For the wider context on the flows themselves, see the complete guide to remittances to Pakistan.
The short answer
Remittances are counter-cyclical to the recipient economy because they respond to need rather than to return. When the home economy deteriorates, need rises, the currency usually weakens so each unit of foreign money goes further, and migrants draw on savings held abroad. A crisis therefore raises both the motive to send and the value of sending.
The evidence
The clearest natural experiment is COVID-19. In early 2020 the World Bank and most forecasters expected remittances to developing countries to fall sharply — migrants were losing jobs in host countries, and the reasoning was straightforward.
The opposite happened. Pakistan recorded a 37% increase in remittances over June to August 2020 against the same three months of 2019, in the middle of a global pandemic. That was not a Pakistani peculiarity; resilience appeared across most major receiving countries, defying essentially every published forecast.
The pattern repeated. Remittances to Pakistan reached $30.8 billion in 2021, and continued through the 2022 floods and successive rounds of currency depreciation — periods in which every other external inflow was under strain. Recorded flows have since risen further, to $41.6 billion in FY26.
Four mechanisms, not one
The usual explanation is that migrants care about their families. True, and insufficient — it explains why remittances exist, not why they rise at precisely the moment the sender is often also worse off. Four distinct forces operate, and they happen to push the same way.
1. Need at home rises
The straightforward altruistic channel. Job losses, inflation, a flood or a medical emergency raise what the household at home requires, and transfers respond. This is real and it is the largest single component, but on its own it would predict that migrants send more even when they cannot afford to — which is not what happens.
2. The exchange rate improves the terms
This one is underrated and mechanical. When the recipient country's currency depreciates, every unit of foreign currency buys more at home. A worker sending a fixed rupee amount to cover a fixed set of household needs can send fewer dollars to achieve it; one sending a fixed dollar amount delivers more.
Which of those dominates depends on why the transfer exists, and the empirical answer in most crises is that senders increase the foreign-currency amount too — because a weaker currency usually arrives alongside inflation that raises the rupee cost of the same basket. The important point is that a depreciation makes remitting more effective, so it lowers the price of solving the problem.
3. Migrants hold savings abroad
A crisis at home is not usually simultaneous with a crisis where the migrant works. That asymmetry is what makes the flow useful. A Pakistani nurse in Manchester or a driver in Dubai has income and often accumulated savings denominated in a currency unaffected by events in Karachi. A one-off transfer out of savings is possible in a way it is not for a domestic relative facing the same shock.
This is why remittances behave like insurance while being nobody's insurance product. The risk has been geographically diversified by the act of migration itself.
4. It is frequently a contract, not a gift
Here the economics gets more interesting than the sentiment allows.
The reference point is Lucas and Stark (1985), who tested pure altruism against household data from Botswana and found it did not fit. Remittance behaviour tracked things altruism alone could not explain — the migrant's inheritance prospects, family asset holdings, the pattern of who had been supported through education.
They proposed tempered altruism, or enlightened self-interest: migration as a household strategy for spreading risk and sharing gains rather than an individual decision. The family invests in sending someone abroad; the migrant remits; the obligation runs in both directions across time. Under that reading, a crisis transfer is not only generosity but the payout leg of an implicit contract — and contracts are honoured in bad times, which is when they are worth having.
The four mechanisms are not competing explanations. In any real transfer they are entangled, and their relative weight differs by corridor, by how recently the migrant left, and by whether they intend to return.
The part most coverage misses
A crisis increase in recorded remittances is not the same as a crisis increase in remittances.
COVID is the cleanest illustration. Border closures did something no policy had managed: they shut down the informal channels. Nobody was hand-carrying cash on a flight home, because there were no flights. Informal value-transfer networks, which depend on movement and on offsetting flows of trade, were badly disrupted. Money that had always been sent had nowhere to go except the banking system — where, for the first time, it was counted.
Part of that 37% is therefore a measurement effect: not new money, but existing money becoming visible.
This is not a reason to discount the resilience finding. Recorded flows still rose during a period when every model predicted a collapse, and the formalisation channel cannot explain all of an increase that size. But it does mean the crisis multiplier is smaller than the headline suggests, and it explains why some of the gain persisted after the crisis passed — flow that moves into banks tends to stay there.
The same logic applies to exchange-rate stabilisations, where a narrowing kerb premium pulls flow into formal channels for entirely different reasons. That case is worked through in how remittances affect the exchange rate.
What this means in practice
For households, this is the finding that matters most in the whole literature: the money is most reliable exactly when it is most needed. No formal insurance market in Pakistan offers a comparable product to a low-income family.
For policymakers, the temptation is to treat remittances as a shock absorber and plan accordingly. That works for domestic shocks and fails completely for the other kind. A Gulf construction slowdown, an oil-price collapse that tightens fiscal space in Riyadh, or a recession in the UK transmits directly into the external account within months, and counter-cyclicality offers nothing — because the shock is on the sending side.
Pakistan's concentration makes this concrete. With roughly 45% of recorded remittances coming from Saudi Arabia and the UAE, the flow is well diversified against Pakistani shocks and barely diversified at all against Gulf ones.
For forecasters, the practical rule is to model remittances against host-country labour market conditions, not domestic ones. Domestic distress raises the flow modestly. Host-country employment determines whether there is anything to raise.
What to watch
- Host-country employment, particularly Gulf construction activity and new visa issuance, which lead Pakistani remittances rather than follow them.
- Whether a surge persists. A genuine income-driven increase decays as the crisis passes; a formalisation-driven one does not, because the flow has permanently changed route.
- The kerb premium, which tells you whether the recorded number is tracking the real one or drifting away from it.
The reason remittances behave differently from every other capital flow is that they are not a capital flow. Nobody is pricing risk or seeking return. They are the financial expression of an obligation that predates the crisis and outlasts it — which is exactly why they show up when everything else is leaving.