Drive cautiously down China’s Belt and Road


By Shahid Yusuf

China’s Belt and Road Initiative (BRI) aims to create a Eurasian economic corridor and a string of economic hubs anchored to Chinese cities,  thereby generating a development dynamic that is advantageous to China’s growth. The investment and trade generated by BRI could enable China to sustain a growth rate of 6 to 7 percent and double its GDP between 2010 and 2021. As of end 2016, $900 billion worth of BRI-related projects were planned or under implementation – with loans and credits from Chinese banks amounting to $1.2 trillion (not all for BRI projects). Chinese agencies claim that the BRI will eventually absorb between $4 trillion and $8 trillion.

But what are the benefits and risks for countries accepting BRI-linked financing to build transport and energy infrastructure?

To this day, the BRI remains a patchwork of projects without a well-articulated strategy backed by solid analysis of the potential benefits for China and countries that will borrow from Chinese entities to finance large infrastructure projects. This is critical if the politically less-than-stable countries in Central and South Asia with a poor track record of sound policymaking are to benefit from BRI. In order to service BRI loans, the investment in transport and energy infrastructures plus any associated technology transfer must attract private investment in tradable goods and services and increase export earnings from exports.[1] Whether such a virtuous spiral of investment and exports will ensue, is far from certain. Moreover, infrastructure building and mining on the scale envisaged could lead to severe environmental degradation absent the enforcement of strict regulations, which are either not in place or enforced with a light touch.

There are other reasons for proceeding cautiously down the Belt and Road. The terms and conditions of loans extended by Chinese entities are less than transparent. Furthermore, the governance and finances of the more than 50 Chinese state-owned enterprises that are responsible for major BRI projects are opaque, and their capacity to manage and implement complex transnational projects is untested. Contractual relations with such entities could prove to be tendentious if projects fail, the quality of work and materials is poor, or if lax environmental standards cause damage. The Tharparkar project in Pakistan is a case in point.

This context elicits the following questions and concerns that deserve closer attention and more systematic study.

Can China finance BRI projects to the tune of several trillion dollars from its own resources? And if not, will China need to tap the international bond market for the bulk of the financing? By doing so, its indebtedness would increase and it would absorb considerable risk associated with lending for long-term projects in countries such as Uzbekistan, Pakistan, Sri Lanka, and Laos. In the end, given the current state of China’s forex reserves, will the outlay on BRI be an affordable but not game changing $25 billion per year?

China’s neighbors worry that the purpose of BRI infrastructure and connectivity is to further Chinese exports and geopolitical ambitions. Many are already on the slippery slope to deindustrialization and BRI could accelerate the process. Existing light consumer manufacturing would be imperiled and the likelihood of diversifying into more complex products would be greatly diminished because of China’s competitive advantage in a wide range of manufactures.

European experience suggests that cross-border transport infrastructure has not led to regional convergence. If anything, it has tended to increase regional disparities by making existing hubs more dominant and disadvantaging nearby regions in the hubs’ shadow. Rail links between Milan and Naples have strengthened hub economies while contributing little to the development of Southern Italy. A study of road infrastructure building in Portugal came to similar negative conclusions: greater accessibility did not improve the cohesion and purchasing power of less developed parts of the country.

To service loans from China and other borrowers, countries on the receiving end of infrastructure investment will need to greatly expand their exports and run trade and current account surpluses. Given recent trends in manufacturing and slower growth of world merchandise trade, is that likely? In 2016, China ran a trade surplus amounting to $250 billion with participants in the BRI. Could countries such as Pakistan (which runs a $13 billion trade deficit with China) possibly narrow and reverse the trade gap and run surpluses with its hyper competitive neighbor?[2] If they do not, what is the return to these countries in the form of long term gains from infrastructure? In other words, how much growth could BRI projects unlock by way of tradable goods and services? Furthermore, if highly indebted countries are unable to repay these loans, what are the consequences for Chinese firms and for their bankers?[3] Taking over assets that will need to be marked down would involve absorbing large losses.

What is the risk of BRI exacerbating the resource curse in countries such as Kazakhstan, Turkmenistan, and Afghanistan? Could the creation of the BRI trade corridor render them even more resource dependent and stunt their non-resource based tradable sectors?

So far, China’s projects in its own Western provinces have at best yielded modest returns. The profitability of China’s foreign direct investment in developing countries has also been low. This suggests that the cross-national infrastructure projects intrinsic to BRI will be costly to build and the financial returns are likely to be meager, at least in the medium term. Political changes in destination countries could easily affect project outcomes. Political risk could discourage participation by investors from developed countries.

Geopolitical issues need to be factored in. China’s actions have alarmed some of its neighbors – India in particular.[4]  Chinese closeness to and support for Pakistan could contribute to continuing friction between Pakistan and India. Political tensions within and among countries, sporadic violence (as in Pakistan’s Baluchistan Province), and arms races in South, Southeast, and East Asia may undermine the BRI – as will continuing discord in the Middle East. How might these developments and others affect growth prospects is a key question.

Shahid Yusuf is Chief Economist of the Growth Dialogue at George Washington University and an adjunct professor at Johns Hopkins University.

Note: The views expressed in this article are the author’s and do not necessarily represent those of Pakistan’s Growth Story.

[1] Premier Li Keqiang referred to technology transfer as China’s, “golden business card”. Financial Times (2017, July 18th p.9).
[2] Between 2006/7 and 2015/16, Pakistan’s exports to China went from $575 billion to $1.63 billion. Meanwhile China’s exports to Pakistan increased from $3.5 billion to $12.1 billion (Source: Figures in the Financial Times indicate that China’s exports amounted to $16.5 billion in 2015.
[3] Down the road, servicing the loans from China will be burdensome for many countries. Chinese firms have already encountered problems with projects in Myanmar, Sri Lanka and Indonesia. Chinese SOEs that are spearheading BRI, such as the China Railway Corporation, are themselves increasingly in debt to Chinese banks – CRC’s debts amount to $558 billion and these are rising largely because much of China’s 22,000 high-speed rail network runs at a loss (Source:,
[4] In response to BRI and disputes along its northern border with China have induced India to launch its own initiative extending from Africa to Southeast Asia variously called the “Spice Route” the “Blue Revolution” and SAGAR – “Security and Growth for all in the Region”. India is also investing $300 million to lease the 2,000 acre tract of land which is the site of the largely deserted Mattala Rajapaksa Airport adjacent to Hambantota Port in Sri Lanka in order to prevent a Chinese takeover of the facility and to control China’s access to the port that it has leased for 99 years (Source:,

Now that provinces have more control over taxes, do they collect more revenue?

income tax pic

By Ghazan Jamal

The 18th Amendment to the Constitution and the 7th National Finance Commission (NFC) award supporting it give the provinces greater control over their fiscal affairs, but has that increased revenue collection? This post compares how Sindh, Punjab, and Khyber Pakhtunkhwa (KP) have raised tax revenue following devolution.

Pakistan’s tax-to-GDP ratio in 2016 stood at 12 percent. According to a recent paper by the IMF, this represents an encouraging increase over the previous three years. As good as that sounds, Pakistan still lags behind other emerging market economies in terms of tax to GDP ratio.[1] Poor tax collection by provinces is an important factor in Pakistan’s overall unsatisfactory tax effort.

The provinces can collect taxes on agriculture, property, automobiles, services, and stamps. However, they continue to depend largely on transfers from the federal government. Based on revised budget estimates for fiscal year (FY) 2015-16, federal tax receipts accounted for 76 percent of the total general revenue in Sindh, 78 percent in Punjab and a staggering 92 percent in KP. Of the total tax revenue in Pakistan, provincial tax collection only accounted for 7.7 percent in 2016 – a meager 0.3 percent of Pakistan’s GDP. This is partially explained by the fact that the two most buoyant avenues of taxation – income tax and general sales tax (GST) on goods – remain with the federal government.

While it appears that devolution has galvanized the provinces to promote social sector development, like in health and education, the increased share of the provinces in the NFC award (as a step towards fiscal decentralization) seems to have limited the incentive of provinces to increase their own revenue generation, since they now receive more funds from the federal government. So what exactly is happening with provincial tax levels since devolution?

Figure 1 below shows the provincial revenue trend for the three provinces starting from FY 2010-11 (the year devolution occurred), till FY 2014-2015, for which the latest actual figures are available.


In FY 2010-11 Punjab managed to collect about PKR 66 billion in provincial taxes while for the same period Sindh collected PKR 42 billion.  The KP government at the time, led by the Awami National Party (ANP), only managed about PKR 3.5 billion. Such low numbers can partly be explained by a two-year tax exemption given by the KP government because of the negative impact of militancy and floods on the business community of KP.

Remarkably, in comparison, Sindh increased its provincial tax revenue to about PKR 96 billion by 2015, a 126 percent increase. However, as explained later, this was primarily achieved through a special tax on imports levied by the Sindh government. On the other hand, KP under Pakistan Tehreek-e-Insaaf (PTI) has been able to take provincial tax receipts to a little over PKR 11 billion by 2015, a 228 percent increase. The big jump came in 2012-13 when provincial tax revenue almost doubled from PKR 3.6 billion in the previous year to PKR 7.4 billion primarily as a result of ending the tax break given two years ago. This was in the ANP government’s last year in office. PTI in its first full year continued with this sharp increase, taking provincial tax revenue to a little over PKR 10 billion. However, between 2014 and 2015 there was a moderate increase of slightly more than PKR 1 billion. Overall, KP still lags behind in revenue collection and continues to depend more heavily on federal transfers compared to the other two provinces.

Similar to the federal level, indirect taxes – specifically general sales tax (GST) on services – remains the main contributor to provincial tax revenue. In 2015, direct taxes only contributed a little over 20 percent in both KP and Punjab. While in Sindh it contributed less than 7 percent of all provincial tax receipts for the province.

In KP, GST on services accounted for 56 percent of the total provincial taxes in 2015. Although KP’s tax on agriculture has increased from PKR 20 million in 2012 to PKR 66 million in 2015, its share remains worryingly low. In the case of Punjab, GST on services contributes about 44 percent of the total provincial tax revenue. Interestingly, stamp duty is also a significant contributor to provincial taxes in Punjab, about 22 percent, unlike in the other two provinces. However, even in Punjab agriculture tax contribution is minuscule, contributing PKR 762 million in 2012 and by 2015 reaching the PKR 1 billion mark, which constitutes barely one percent of the provincial tax revenue for that year.

Similar to KP and Punjab, in Sindh GST on services is the largest contributor to provincial tax, contributing about 52 percent. However, the Sindh government, taking advantage of being Pakistan’s main port of entry for goods, charges a Sindh Development and Maintenance of Infra-Structure cess on imports coming into Sindh by air or sea. Its contribution is another 26 percent to the province’s tax revenue in 2015. Other provinces have raised concerns over this form of taxation pointing towards Sindh taxing goods that are meant for other provinces. The share of tax from agriculture is even lower in comparison to the other two provinces; it stands at a mere 0.3 percent of the total tax revenue.

Contrary to popular belief, the absolute Rupee value of tax collection by the provinces actually increased in recent years. However, the provinces still only contribute less than 8 percent of all tax revenue. This means there is still a lot of room for all three provinces to improve their tax collection. However, care must be taken not to implement taxes that distort the cost structure of firms (as in the case of Sindh’s cess, and the provincial tax on internet services) and are growth inhibiting and/or increase the burden of income taxation on those who already pay taxes (as in the case of Punjab’s tax on some services). While improving their tax effort, provinces must also avoid taxes that are regressive, as seen in the tax on cellular phone services.

Ghazan Jamal is a Pakistan Country Economist at the International Growth Centre. 

What the data from 2014 reveals about income tax, Part 1: Tax filers and non-filers

moneyAnjum Nasim and Umbreen Fatima

The government of Pakistan collects income tax from both filers and non-filers of income tax returns.  In this article, we examine the share of tax collected from filers and non-filers in total federal income tax in tax year 2014 (TY2014).[1] We also comment on the limitations of withholding taxes if the burden of taxes is to be shared equitably across tax filers and non-filers.[2]

The income tax returns filed in TY2014 by companies, association of persons and individuals were in excess of 0.856 million.[3] Tax filers contributed about 63 percent to total federal income tax. The remaining 37 percent was collected from non-filers in the form of withholding taxes.[4]

How income tax revenue breaks down by mode of collection

In TY2014, gross federal income tax collection was Rs 919 billion, of which Rs 64 billion was refunded to taxpayers. Net income tax collection constituted 36 percent of federal tax revenue.

There are three principal modes of federal income tax collection in Pakistan: withholding taxes, voluntary payments by taxpayers and collection “out of demand”: taxes raised by the Federal Board of Revenue (FBR) on the basis of tax audits and from the recovery of outstanding payments.

As seen in the chart below, withholding taxes contributed the most to gross income tax collection (62 percent), followed by voluntary payments (29 percent), and collection “out of demand” (9 percent).[5] These ratios were 51.5 percent, 37.3 percent and 11.2 percent a decade earlier.

Distribution of income tax by mode of collection

(source: FBR Year Book 2014-2015)

Withholding taxes have been instrumental in increasing the share of income tax in total federal taxes from 14.4 percent in 1989/90 to 37.9 percent in 2014/15.[6] However, it is important to note that some withholding taxes may not be different from indirect taxes – such as customs and excises – thus exaggerating the share of income tax in total federal taxes.

The collection “out of demand” has been very erratic over the years. From 2001 to 2014, it was as low as 3.2 percent of gross income tax collection in 2006/07 and peaked at 17.6 percent in 2009/10, with considerable variation over the period.

The share of tax filers and non-filers in income tax

The total income tax declared, as listed in the TD2014, was Rs 491 billion.[7] If we assume that the taxes ‘out of demand’ (Rs 81 billion) and those listed under the category of ‘miscellaneous’ (Rs 4 billion) were raised exclusively from tax filers in 2014, then the tax paid by tax filers in TY2014 was Rs 575 billion. Since gross income tax collection was Rs 919 billion, it follows that Rs 344 billion of gross income tax collection, exclusively in the form of withholding taxes, is attributable to non-filers. The distribution of taxes paid between filers and non-filers is shown below.

Distribution of income tax between tax filers and non-filers

(Authors’ calculations)

We next consider withholding tax collection, the main component of gross income tax collection for TY2014. The total withholding income tax collection in TY2014 was Rs 572 billion. Subtracting our estimate of Rs 344 billion as withholding tax payment by non-filers, yields Rs 228 billion as withholding tax payment by tax filers. The figure below shows the relative distribution of withholding taxes between filers and non-filers:

Distribution of withholding income tax between tax filers and non-filers

(Authors’ calculations)

While the withholding tax paid by tax filers was Rs 228 billion, other components of their income tax were voluntary payments (Rs 263 billion), tax “out of demand” (Rs 81 billion) and miscellaneous payments (Rs 4 billion). The figure below depicts the distribution of income tax paid by tax filers between these components. On the aggregate, tax filers paid 39 percent of their income taxes in the form of withholding taxes. This ratio could, however, vary across different categories of tax filers (companies, association of persons, and individuals) and within the same category of tax filers. On the other hand, all non-filers pay 100 percent of their taxes in the form of withholding taxes.

Distribution of income tax for tax-filers

(Authors’ calculations)

Government efforts to increase the tax base

Although random audits of tax filers generated 7.4 percent of income tax collection in TY2014, FBR has not been very successful in recovering unpaid income taxes from non-filers. A project, ‘Broadening of the Tax Base’ that started at FBR in 2013, added 130,000 new tax filers, who filed only Rs 1.8 billion worth of tax returns. An additional tax demand by Rs 36 billion was raised but less than Rs 1 billion was collected as of 2016.[8]

Recently, the government has made it harder for non-filers to remain outside the tax net by imposing higher withholding tax rates. The cost of remaining outside the formal system can be increased further by expanding the withholding tax net and widening the difference in withholding tax rates between non-filers and filers. But even though this creates an incentive for non-filers to file their tax returns, it may not raise income tax collection if non-filers join the ranks of filers but declare their incomes to be below the threshold level of income required to impose an income tax. In this case, there would be no increase in taxes that are not withholding taxes, but the withholding tax revenue would decrease because lower rates would apply to previous non-filers turned filers.

The effort to broaden the tax base should also be intensified by recovering unpaid taxes. Identifying new tax filers and recovering tax from those who have assiduously avoided tax payment can be very challenging. It requires coordination and information sharing across ministries and departments at the federal and provincial levels, rigorous investigative work and data analysis at the FBR, and a robust legal team. Most importantly it requires strong political backing at the highest level of government to take on the economically powerful but rogue elements of society.

Without such a commitment, taxes will continue to be raised through greater indirect taxation or squeezing more taxes from existing tax filers. This will have even graver long-term consequences for social cohesion and the writ of the government.

The analysis in this article is based on Tax Directory for 2014 (provisional edition) (TD2014) and the FBR yearbook 2014-15, both available on the FBR website.

[2] Withholding income taxes are taxes that are collected at source, e.g., by employers at the time of payment of wages and salaries or by companies at the time of payment of dividends. Some withholding taxes in Pakistan are presumptive income taxes, e.g., on electricity and phone bills, or on cash withdrawals from banks, or on imports and exports. Some withholding taxes are adjustable against final tax liability but others are un-adjustable.

[3] We calculated the number of tax filers from TD2014 to be 0.856 million. TD2014 clarifies that a “considerable number of manually filed Returns could not be entered in the system […] due to missing identifiers on the Returns”. If account is taken of the manually filed returns, which have not been entered, the number of tax filers will exceed 0.856 million.

[4] Since our estimates of the shares of tax filers and non-filers in income tax are based on TD2014, the omission of a number of manually filed returns in TD2014 may underestimate the tax share of tax filers and over-estimate the share of non-filers.

[5] Source: FBR Year Book 2014-15, Revenue Division, Ministry of Finance, Government of Pakistan (available at:

[6] Calculations based on data in ‘Table-1: Federal Tax Receipts’ in FBR Year Book 2014-15.

[7] While the FBR Year Book provides figures for gross income tax collection, the Tax Directory provides the list of all tax filers (companies, association of persons, and individuals) who file their tax returns, as well as the taxes paid by them i.e. the sum of their withholding taxes and voluntary payments.

[8] See Chapter 3 in Revenue Division Year Book 2015-16, FBR, Ministry of Finance.

Anjum Nasim is a senior research fellow at the Institute of Development and Economic Alternatives (IDEAS).

Umbreen Fatima is a research associate with IDEAS.