 Thank you for the invitation. Thank you for being here. I'm presenting work with Abrams Tagum at Wider There are two papers Embedded in here, but I'm not going to take 30 minutes. I'm determined to essentially use the first Present what we found as motivation for the second and focus more on the second And how does it work when you did it and not when I did it? So Just to give you a preview or a map of the talk Just in case I run out of time so what we're interested in is the relationship between taxation and State capacity are more specifically two Indicators of state capacity one is tax capacity and that's a measure we developed in the first paper as part of the wider fiscal states program and The second is accountability And I'll elaborate on precisely the accountability measure and it's motivated by three observations or by an observation of three things in Sub-Saharan Africa since about 2000 on Average and in most countries the tax to GDP ratios have increased tax capacity has increased and Accountability has increased So what's the relationship between them? That's what we're interested in Tax capacity are the change in improvement in tax capacity is not explained by accountability but accountability and specifically Vertical accountability which is a measure of the quality of the electoral process and party competition that has been increased by the increases in tax GDP increases in tax GDP have led to or contributed to improvements in vertical accountability so there's a positive effect of taxation on Indicators of state capacity So as I said, it's based on two papers. The first is really developing our measure of tax capacity and the determinants of that tax capacity and That's come from the wider project that Antonio and Kunal had and That paper has been published in the special issue of the Journal of Institution of Economics The core points about that paper is first economic structure or variables that capture the economic structure is the most robust determinants of the tax to GDP ratio Various institutional or political or governance variables can be included, but they're not robustly or consistently significant So we use that and anybody who is at the tax effort session The underlying notion is very similar that you try to you start by trying to get a measure of the ratio of the potential Tax revenue that could be collected given the structure of the economy and the actual tax revenue We go a step further and our measure of tax capacity is the trend component of That tax effort or collection efficiency variable and the importance about the trend component the underlying trend is you're taking out all of these Random or unobserved factors that can cause tax GDP ratios to go up go down to change from year to year And introduce an awful lot of variability When you're trying to estimate, you know the ratio of potential to actual because there's lots of things that will happen that will affect actual tax collection So we take those out and look at what's the underlying trend and We treat that as the measure of the trend in performance And we find that that's increasing over time And then we we've calculated that for each country each of the Africa for a sample of about 34 35 African countries From 1980 through to about 2018 and then we try to say well, okay, can we identify? Variables are factors that explain the increase in tax capacity and There are four variables that are robustly significant are the best to explanatory variables and one is it increases with private consumption and that makes sense Because the private consumption component of GDP is a good measure of what's happening to people's incomes So if their incomes are rising they spend more They pay more income tax. They pay more sales taxes So it's pretty that's pretty intuitively logical Perhaps more surprisingly It also increases with resource rents But there's more heterogeneity in the countries there and I'll partly illustrate that but we do find it's a positive association You might have expected it to be negative And it also increases with the equal distribution of resources, which is a measure of how equitably Public spending or public services are distributed and again, that's intuitively plausible because if people think that They're getting a more that the government is being fairer are more equitable in the way. It's delivering public services. They should be Prepared to pay a bit more tax It decreases with corruption and again you'd expect that and actually that might partly explain why the resource rents It's positive because in a way what you're saying is you've got the if you've got the if Resources are actually feeding corruption Then the adverse effect of corruption can offset the potentially beneficial effect of having resources and Then fundamentally we do in that work accountability is not one of the drivers of capacity So having established that capacity is increasing we know the variables that explain it We then want to turn to and that accountability is not a direct Driver of the improvements in tax capacity We then ask the question. Well, okay. Does taxation have an effect on accountability? There are lots of theoretical or conceptual Reasons to think it would the basic idea is if the median voter taxpayer if they're paying more tax they will demand More accountability are more responsiveness for the government. That's the underlying conceptual motivation So that's what we turn to and that's what I'll focus on Two sources of data are used in both of the papers all of the Institutional or governance variables that we use they all come from varieties of democracy and in particular the vertical accountability Core measure which captures the ability to kind of freely organize in political parties Participate in free and fair elections. So it's the quality of the electoral process We also look at other measures of accountability and demonstrate that Taxation has no effect on the other measures of accountability I'll come to that later and then the revenue variables are all from the government revenue data set and it's total government tax and Total government revenue We do some work distinguishing between direct and indirect taxes and we control for non-tax revenues our resource revenues It's just to illustrate some of the the data the black line on the top left Diagram that's just illustrating the increase in tax revenue and that the dotted line above it is total revenue and That the main point there is from the late 1990s you see a fairly steady increase The second panel on the right at the top that's looking at the measure of tax capacity and again The high values at the start are partly driven by data So I don't interpret those but again you see from the late 1990s a steady increase in Tax capacity now the fact that the value is above one does not mean and should not be Interpreted as saying that these countries are collecting more than their potential. That's not what it captures But what it is saying is that there is an underlying improvement in The tax capacity and their ability to collect What what the potential is that it is that the underlying drivers are improvements The bottom diagrams which I won't go into just show that there is There isn't an awful lot of variability over time for the low-income countries and The improvement in tax capacity and revenues is pretty consistent and steady for the low-income countries There's a lot more Variability in what's happening with the middle-income countries Many of which of course are also the resource rich countries. So there's a lot more heterogeneity Within countries and over time for those but we control for those factors in our estimation And then this is just showing how Vertical accountability has been improving in sub-Saharan Africa over the decades There are still some countries that have very low values But you can see that the average has been increasing and the Variation around that average has been compressed so the countries they are getting better and Some of the the countries with worse accountability are getting better faster than some of those with higher Accountability, so they're all on average with with some exceptions. They're all converging towards a higher level of vertical accountability more closely clustered around the mean So you're getting this steady increase so taxes are increasing tax capacities increasing Accountability is increasing This is just some simple scatter plots to start with so what you find is broadly speaking there is a positive unconditional Correlation between accountability and total tax There's no relationship on average between Accountability and non-tax revenue Non-tax revenue doesn't matter The relationship between accountability and tax holds for direct taxes and for indirect taxes in total now we discuss in the paper you know there's a lot of literature saying that well The pressure from taxpayers to get governments to be responsive should be greater for direct taxes and indirect taxes because they're more visible We're working at a country level Across country level. So my argument is that what particularly at the country level and to some extent if you think about the individual It's whether they feel they're paying more tax that matters It doesn't really matter if they think well, I'm paying more Sales tax or I'm paying more personal income tax. The bottom line is am I paying more tax? Has tax on fuel gone up in which case I'm paying more for petrol has tax on clothes gone up Or has my income tax gone so really what they're concerned about is am I paying more tax or do I think I'm paying more tax and On that basis, it's quite reasonable to use the tax GDP ratio as your proxy Now Okay, for those who know the economy who are no econometrics This is the technical bit for those who don't you can close your eyes or turn off. I just try to play explain the intuition the basic model is a fixed effects model so fixed effects means that we Have a country fixed effect that accounts for unobserved country characteristics Time fixed effect that accounts for something that might be happening in particular years the tax variable and then a number of control variables there in that large X and That's the basic model, but we also Estimate that with a lag dependent variable in other words. We have the lag value of accountability The easiest way to interpret that is we're looking the first equation is looking at what's happening in the long run The second one is looking more at what's happening in the short run and you can interpret it as does tax explain the changes in accountability and then Anybody who does econometrics know you have to have an identification strategy. You have to allow for endogeneity So so intuitively what you're trying to do here is you want to reassure yourself Or convince yourself or in practice convince the referee that the effect you're finding is Genuinely an effect of tax on accountability and not caused by some other factor That means that they happen to go together But it's not because tax influences accountability is because something else is driving both of them To move in the same direction our main instrument for that is Actually built on work that I did with Christian and others is we use measures of economic shocks to instrument for taxation because Economic shocks do have a significant effect on the amount of Tax revenue that can be collected, but there's no reason why they should have an effect on accountability We also use it and another technical instrument There is one problem. You might say ah, but what if the economic shock coincided with an election and as a result of the loss of income associated with the shock the government Interfered more in the electoral in the election And that might sound plausible that should mean that the shock will be associated with a reduction in Accountability because that would be picked up in the electoral accountability measure We test for that no effect We account for presidential elections doesn't alter the results Now I can quite go go quickly through it and just say Basic results Don't worry about the fact of the coefficient changes It always does when you alter the estimator. The point is the effective tax is consistently positive and significant Whatever we do then we test for the Indigeneity various ways again the effective tax is positive and significant whatever we do incidentally aid The coefficient on aid is always positive, but it's not sorry. It's not always significant I just think that's an interesting thing to put in especially because so much of work at wider particularly with Finn Was on the aid? Well look aid seems to be having a positive effect as well So don't worry about all these people that say aid is bad in sub-Saharan Africa. It's not No effect whatsoever on the other measures of accountability Diagonal accountability takes into account civil society the media free media horizontal accountability is the Relationships between different levels of government and checks and balances Judicial accountability is the quality of the judiciary no effect One of the reasons for that is the final point I want to emphasize is you look at the coefficients on the lag variables They're very high for the other measures of accountability. What that's telling you is they don't change much They're not changing very much over time So there's not how you're going to explain if there's not much change over time You're not going to be able to explain the change over time Vertical accountability is the one that changes the most It has the lowest value on the line So there is more change and variability to explain and that's what tax is picking up Final bit we look at other factors that might determine it Tax remains significant and positive Now what seems to be very important are these Neopatrimonialism index which is also from VDM and a critical media having a critical media a free press is good Having a neopatrimonial state is bad Now they have higher coefficients the way I would interpret that is there More important to understand the level of accountability Whereas Tax is less important for the level. It's still important But our focus is on what's the effect of tax on the change? so to conclude taxation is positively linked with with accountability and Basically tax explains about 10% of the increase in observed vertical accountability Now that may sound small But it's still Given the level that we're working out with aggregate variables proxy variables It's a robust Effect So it allows us to conclude that taxation and it's true for both direct taxation and indirect taxes It does have a beneficial effect on accountability If you look at individual sales or trade taxes, no you don't find the effect And if you look at resource or non-tax revenue, you don't find the effect. Thank you We went over a bit but