The budget season is an important occasion to raise questions about the nuts and bolts of fiscal policy. One such nut – subsidy – has become harder to crack this year because of heightened external and domestic inflationary pressure. The government can use every taka it gets to soften passing through the global price increase and exchange rate depreciation to administered prices of electricity, gas, diesel, and fertiliser.
Where can it save some?
Subsidy on remittances is one among several that deserve to be looked at. It came in July 2019 at 2% and expanded last January to 2.5% on every remittance dollar. Remittances, like most other sources of income, have no microeconomic externality. Hence, no efficiency rationale for subsidising it. The beneficiaries predominantly are neither the bottom nor the upper 20%. An argument could still be made on socio-political grounds ("remittance soldiers") but was not.
The intention was to incentivise the propensity to remit through formal channels. Now that we can look back at the experience, two questions need answers to make a prudent choice on whether to continue or rollback this subsidy. First, what has the subsidy changed in the market for remittance dollars? A corollary, was the intended shift towards formal channel achieved? Second, do the changed circumstances justify continuing it?
Subsidy spilled over to the informal rate
Remittance flows through formal channels in the post subsidy period followed the same pattern as they did before the subsidy. Remittances increased $2.2 billion in FY18 and $1.5 billion in FY19, the two years preceding the subsidy. In FY20, the first year of the subsidy, remittances increased $1.8 billion followed by a Covid-induced acceleration of $6.6 billion in FY21. In the first ten months through April FY22, when the subsidy increased, remittances declined by $3.3 billion relative to the same period last year. There is nothing definitive one can conclude about the effect of the subsidy from these data.
If we think a little harder, a simple supply and demand story can explain what difference the subsidy made. Differences in motivation on the demand side and regulatory reach divide the formal and informal markets into interdependent segments. The informal market, being outside the purview of the regulator, provides what the formal market does not – unquestioned access to dollars. There are many players on the supply and demand side in both markets where information on exchange rates and public policy is common knowledge.
These conditions make exchange rates in the two segments move together with a premium on the informal rate. The premium compensates the costs and risks of trading in the informal market. It is smaller the more easily dollars travel from one segment to the other. When the informal rate exceeds the formal rate plus the premium, the supply of dollars in the informal market rises, leading to a fall in the informal rate. When the informal rate is lower than the formal rate plus risk premium, dollar supply dries, leading to a rise in the informal rate. Arbitrage makes sure the informal rate equals the formal rate plus the risk premium in equilibrium.
Such a parity between the two rates implies a subsidy targeted at the formal rate must spill over to the informal rate. Remittance is a significant source of supply of dollars in both segments. As remitters respond to the increased attractiveness of the formal channel due to the subsidy, supply in the informal segment falls until the informal rate increases enough to compensate the subsidy. The difference between the formal and informal rate converges to the subsidy plus the pre-existing informal rate premium. The informal rate will depreciate more than the formal rate to reach such a convergence when both are under depreciation pressure due to common factors, as happened in the last 15 months or so.
The data on informal market premium fits closely with what theory predicts. I am taking the cash dollar market as a proxy for the informal market. The average premium during the twelve months preceding the subsidy was 0.8%. This difference between the informal and formal rates rose to 2.1% during the twelve months after the subsidy. Covid shocked the difference down to 0.5% in July 2020 as trade dried in the informal market. It recovered rather quickly from August onwards, shooting up to 2.4% in 2021, and further to 2.7% last April following the enhancement of the subsidy. Since the introduction of the subsidy until April this year, the formal market rate depreciated 2.6% and the informal rate 4.6%.
The stylized evidence on spillover is hard to refute. It appears large enough to significantly neutralise the diversion effect of the subsidy from informal to formal channels.
Incentivised all with mild volume effects
The subsidy certainly increased the income of all remittance recipients irrespective of channels. Those taking the formal route get the subsidy and those using the informal get almost an equally higher rate. The subsidy thus has the effect of redistributing cash from the government and informal market buyers of cash dollars to the remittance recipients.
The redistribution amounted to around Tk4,300 crores in FY21. Assuming (net) one crore workers abroad, this means a transfer of Tk4,300 per worker, over 1.7 times the gift from the Prime Minister last year to families hit hard by the second wave of the pandemic.
This calculation is based on formal remittance data. Note FY21, the second year of the pandemic, saw a major reverse diversion to formal channels. It worked like a natural experiment, revealing the total remitted through both channels in a normal year. Informal remittances appear at least 40% as large as formal remittances.
By incentivizing all remittances, the subsidy could also have impacted total remittances (formal plus informal). There is some evidence, based on international data, that remittances and exchange rates move together causally. The nature of this relation depends on the behaviour of the regulator. In a stabilised exchange rate regime such as being pursued by the BB in Bangladesh, causation from the nominal exchange rate to the remittance volume is most likely a one-way street.
Arguably then the subsidy in the formal market and the consequent increase in informal rates should boost remittances without any reverse effect. The size of the total boost, given the less than Tk2 per dollar increase in rates in the twelve months ending April this year, was probably not huge (about $300 million, equivalent to Tk2,600 crore, based on a responsiveness parameter estimated by the World Bank in 2012). The government, however, had to spend Tk3,700 crore to induce the increase!
The underwhelming effect of the nominal exchange rate on the propensity to remit is remarkably evident in the long run parameters of interest as well. During the decade ending FY22, remittances increased 61% while stock of workers abroad rose 74%. The difference between the two constitutes growth in remittance per worker. This, presumably the exchange rate-sensitive part of remittance, declined 7% when, in the same decade, the informal rate depreciated 16% and the formal rate 8.4%. Other factors related to conditions in host nations and home appear to weigh much heavier.
The redistribution of cash from the government and informal market buyers to the remittance recipients was not explicitly intended. Unintended does not necessarily equate with undesirable. The policymakers and the public may find this as acceptable as the diversion objective. However, times currently are better for remittance recipients than their neighbours next door or village with no family members working abroad.
The exchange rate has increased substantially in the last three weeks, over Tk10 per dollar, in both segments. They are likely to remain at this elevated level until the shocks are reversed. Higher oil prices may tighten labour markets, with positive effects on wages, in oil exporting countries where the bulk of Bangladesh's overseas workers are. These will enhance the income of the recipients by boosting remittance inflows. Total remittance is poised to be boosted any way from the 617,000 workers who found employment abroad in 2021, another 426,000 in January-April 2022, and counting.
How much additional remittances per annum can be expected from the increased stock of migrant labour, change in exchange rate and the change in oil price? Based on the parameters of responsiveness to each of these estimated in the WB study mentioned earlier, an increase of one million in the stock of migrant workers could add $1.5 billion, a sustained Tk7.5 per dollar higher exchange rate $1 billion and $50 per barrel higher oil price $500 million. These three add up to $3 billion, a ballpark estimate.
Removal of the subsidy will slash 2.5 percentage points from the over 10% depreciation of the nominal rate, thus still maintaining a sizable 7.5% stimulus in one go. It may save the government about Tk5,000 crores, assuming remittances reach $21 billion in FY23. The government could use this to transfer Tk2,500 per capita to nearly 20 million people in the bottom 20% of the population. The remittance recipients would still gain Tk9,500 per worker, higher than the minimum garment wage, if the exchange rates remain at the Tk95 per dollar level.
The power of subsidy as a remittance incentive is watered down when it is on top of such a large exchange rate depreciation. After all, at the individual level, remittance is bound from above by savings which depends on how much the overseas workers make and how much it costs to live where they live. Too much incentivisation may even induce them to reduce the amount of dollars remitted because the taka received per dollar is now much higher. Less buys more.
Zahid Hussain is the former lead economist at the World Bank's Dhaka office.